The GTMnow Newsletter (by GTMfund)

MD files in Claude Code: How GTM teams are giving Claude a brain

GTMnow — Fri, 24 Apr 2026 12:34:58 GMT

Claude Code’s value across teams massively increases with a persistent context layer: a set of markdown files Claude reads before any prompt. ICP, positioning, campaigns, competitive context — all preloaded and shared across the team.

So Claude doesn’t start from scratch. It starts from context for entire team. That shift turns Claude from a tool into a compounding GTM system.

This edition breaks down how to build that system and what to include in your “Claude brain.”

Sales today still relies on manual work: stitching together “Frankenstacks” and toggling between a dozen tabs just to log a single call.

Reevo is breaking the mold. As a vertically integrated Revenue Operating System, Reevo replaces the fragmented maze with a single, AI-native workspace. It captures 100% of your first-party data, from emails to meeting recordings, to help reps find high-intent buyers, automate administrative “garbage,” and generate context-rich outreach that actually sounds human.

No more wasting time on data entry or disconnected tools. It’s one unified engine for the entire revenue cycle. Go Stackless. Learn more at reevo.ai.

Influ2 is hosting a webinar on Tuesday, May 6th, 12–1pm ET on a problem most ABM teams struggle with, but few have solved: every buyer takes a different path to a decision, yet most programs still treat everyone at an account the same way.

Join Influ2 and guest speakers Devin Reed (CEO of The Reeder) and Ed Vander Bush (Senior Manager of ABM at LiveRamp), as they break down the failure points that erode buyer momentum and how Influ2’s newest release, built around contact-level orchestration, overcomes them.

Register here for the session - it’s time to orchestrate around how buyers actually buy.

The forgetting problem

Claude Code is exceptional at execution. Give it a clear task, point it at the right inputs, and it works through problems with a level of rigor that used to require a strong analyst or engineer.

But it has a structural limitation most teams hit quickly: it doesn’t remember anything once the session ends.

With a single folder, written in plain English, you can transform your Claude usage.

The gap between ad-hoc prompting and a structured context layer

CLAUDE.md

CLAUDE.md is where this all starts.

It’s a simple markdown file that Claude Code automatically reads at the beginning of every session. Not a prompt. Not a system message. It’s a persistent instruction layer that loads your GTM context before you type a single word.

The easiest way to think about it: it’s the difference between onboarding a consultant who already understands your business vs. briefing one from scratch every morning. Same capability, very different output.

This idea caught on quickly. Tools across the ecosystem adopted similar patterns: Google’s Gemini CLI introduced GEMINI.md, and Cursor introduced .cursorrules. Different names, same underlying concept: give the model persistent context upfront.

The key is to unlock CLAUDE.md as the entry point to your GTM operating system - the place where core context lives, gets updated, and compounds over time.

The full GTM repository: CLAUDE.md as index, context files as strategy, skills as repeatable execution

The GTM repository: More than 1 file

CLAUDE.md is the entry point. But the real advantage shows up in what CLAUDE.md points to.

The highest-performing teams aren’t stopping at a single file. They’re building a GTM repository — a structured set of markdown files that encode their go-to-market strategy in a way Claude can read, reference, and act on. Instead of re-explaining context or rewriting processes, they’re systematizing it once and reusing it everywhere.

Most of these repositories break down into three layers:

1) Context: your GTM strategy layer

This is the foundation. It’s the knowledge base Claude uses to understand your business — what you sell, who you sell to, how you win, and what matters right now. It doesn’t change daily, but it evolves over time as your strategy sharpens.

In practice, this includes your company narrative, ICP definitions, messaging and positioning, competitive context, and a running library of high-signal insights. When this layer is strong, every output starts from alignment instead of guesswork.

2) Skills: your execution layer

Skills are where teams start to see step-function gains.

A skill is a reusable markdown file (typically stored in .claude/skills/) that encodes how a task gets done — the process, the structure, and the expected output. Unlike prompts, which disappear at the end of a session, skills persist. Once defined, they can be applied consistently across the entire team.

This is what turns one-off outputs into repeatable systems. Content audits, campaign analysis, meeting prep, post-call summaries — all standardized, all reusable.

The impact is measurable. B2B marketing teams report cutting time spent on content audits and campaign analysis by up to 75% after implementing a proper skill layer. Sales teams see similar gains, with meeting prep and follow-up workflows reducing call-related admin by more than two-thirds.

3) Workflows: your process layer

If skills define how to execute a task, workflows define when and why it happens.

These are your core operating processes translated into something Claude can follow — outbound playbooks, content production systems, deal review frameworks, QBR prep, onboarding checklists. It’s the way your team actually works, made explicit and executable.

Every workflow you encode here is a form of leverage. It’s a process your team doesn’t have to rethink — and a process your competitors likely haven’t systematized.

The six files that capture your GTM strategy in a form Claude can act on

The 6 files that give Claude Code a brain

1. CLAUDE.md: The master index

This file is the first thing Claude reads every session. It should be scannable in under two minutes. Its job is to orient Claude to your current reality: what week it is, which campaigns are live, and what is the most important thing Claude should know before any session this week.

The failure mode most teams hit is making CLAUDE.md too long. It becomes a second context file instead of an index. The rule is simple: if something needs more than three sentences, it belongs in a dedicated context file that CLAUDE.md points to.

Update this file every Monday morning. Five minutes. It is the only file that needs to change every week, and that five-minute investment keeps the entire repository oriented to your current state.

2. profile.md: Your company foundation

Product description, use cases, deal profiles by segment, reference customer names, and team structure. This is the foundation every other file builds on. When Claude generates a research brief, writes a sequence, or drafts a case study, it reads this file first.

3. icp-definition.md: Buyer precision

This is where most teams reveal they do not actually have an ICP. Writing it down forces the precision that verbal definitions allow you to avoid. Not ‘Series B SaaS companies.’ That is a category, not a definition.

A real ICP definition includes the employee range and the reason for it, technographic signals that indicate fit, organizational signals (does this function exist?), explicit anti-ICP exclusions, and a qualification framework with must-haves separated from red flags.

One pattern consistent across every high-performing team: they maintain an ICP evolution log. What changed? When did it change? Why? After a year, that log tells you more about your market than the current definition does.

4. positioning.md: Your messaging clarity

Core value pillars, what not to say, competitive differentiation, and objection frameworks. Every piece of content Claude generates, every sequence it writes, and every battlecard it builds, draws from this file. Without it, Claude defaults to generic category language.

The most valuable line in any positioning.md is the list of things you do not say. Competitors’ terms you avoid. Category positioning you want to escape. Adjectives that have lost their meaning. Claude follows these rules with perfect consistency in a way that no human writer ever manages.

5. competitor-radar.md: Living Intelligence

Win/loss patterns by segment. Common objections they raise. Where you beat them. Where you lose and why. This file is updated after every deal, which means it compounds. The team that has been running this file for six months has a competitive intelligence layer that no static battlecard can match.

Claude Code can automate much of the maintenance. One practical workflow: point Claude at a competitor’s pricing page every Monday and ask it to compare this week’s version against last week’s saved version. If anything changed, write a summary and save it to competitive-intel/pricing-changes.md. Stale context is almost worse than no context. It generates confident but wrong output.

6. signal-library.md: Your Intent Engine

Signals with scoring, detection methods, hook angles that convert, and a performance log per campaign. This is the file that turns Claude Code from a content tool into a revenue tool.

The signal library starts as a hypothesis document. After three or four campaigns, it becomes a learning system. Which signals generated the highest reply rates? Which hook angles drove responses? Which touches actually resulted in meetings? The performance log is the step most teams skip, and it is exactly why their signal library never becomes more than a list of guesses.

How skills compound on context: each layer makes the one above it sharper

Claude Code reached $2.5 billion in annualized revenue by February 2026, with weekly active users doubling in the same period. It overtook GitHub Copilot as the most-used AI coding tool within eight months of launch, and 70% of Fortune 100 companies now use Claude in some form - wild.

How to keep the brain sharp

The most common failure mode teams hit after building a repository is letting it go stale. Context files that do not get updated stop being strategic assets and start being noise sources. Stale context is worse than no context: it produces confident but wrong output.

The teams seeing durable results treat repository maintenance as a recurring operational practice, not a one-time setup task.

Less than 2 hours a week keeps the GTM brain sharp, self-correcting, and compounding

The total ongoing time investment is under two hours a week. The question is whether those two hours get spent or not. Teams that maintain the repository have a self-correcting system. Teams that do not, are running on stale context, which is almost worse than no context at all.

How to build yours this week

The practical starting point is not the most sophisticated architecture. It is the minimum viable repository that actually gets used.

Day 1-2: Write your CLAUDE.md and profile.md

Do not start from a blank page. Ask Claude to interview you about your company, your product, your customers, and your current priorities. Use the output as a draft. Spend 30 minutes refining it. This is the foundation that everything else references.

Day 3-4: Write your icp-definition.md and positioning.md

Force yourself to be specific. Concrete employee ranges. Named technographic signals. An explicit list of anti-ICP exclusions. Things you do not say. If you cannot write these with precision, your GTM motion is running on assumptions that need to be tested anyway. Writing the file forces the reckoning.

Day 5: Build your first skill

Choose the task your team repeats most often. Account research before outreach calls. Post-call summaries. Content briefs. Write a SKILL.md that encodes exactly how you want that task done: the format, the tone, the sources to reference, the output structure. Invoke it with /skill-name. Refine based on output quality. This is the proof of concept that gets the rest of the team bought in.

Week 2: Connect to your live data

MCP servers are now available for most CRMs, analytics platforms, and sending tools. Start with read-only access. Test queries against live data. The first time Claude surfaces a genuinely useful insight from your live CRM data without a manual data export, the value of the full stack becomes clear to everyone watching.

Distribution is still the moat

There is a version of this story that sounds purely technical. Markdown files. Directory structures. Skill invocation syntax. That framing misses the point.

What GTM teams are actually building when they build a repository is a distribution advantage. Every skill file is a process your competitor has not yet made machine-executable. Every signal performance log is a compounding learning loop your competitor is still running manually. Every ICP evolution entry is a market insight that your competitor is either guessing at or not tracking at all.

When Claude Cowork launched in January 2026, the GTMnow community described it as the moment Claude Code became accessible to non-technical GTM operators. That was true. But accessibility was not the insight. The insight was the GTM lesson underneath the Cowork story: when software creation compresses, distribution becomes the constraint.

The same principle applies here. When AI execution compresses, context becomes the constraint. The teams that own the context layer own the output advantage. The teams that do not are prompting from scratch every morning, wondering why their Claude code output does not match their strategic reality.

Your CLAUDE.md is a competitive moat in markdown form.

Tag @GTMnow so we can see your takeaways and help amplify them.

SpaceX secured an option to acquire Cursor at a $60B valuation — this reframes how the market is pricing AI tools that drive step-function productivity gains. The infrastructure layer of the AI economy is getting consolidated early.

The GTM Repository guide from GTM Strategist goes deep on the ICP evolution log pattern and the signal library maintenance loop. If you are serious about building this over the long term, the section on automating the update loop is worth the read.

GTM 187: How Figma Scaled From $2M to IPO | Kyle Parrish (First Sales Hire)

Listen through the links in the page above or by searching wherever you get your podcasts “The GTMnow Podcast.”

Monk - just raised a $25M Series A (Footwork + Acrew Capital) to fix B2B collections, the $3T sitting in U.S. accounts receivable that takes 45-90 days to actually land. Most AR tools automate the easy stuff, but Monk is built for the edge cases. Customers are seeing 40% lower DSO and 25+ hours saved per month.

SyncGTM - is building the definitive resource for GTM teams running Claude Code in production. Their benchmarks of GTM skill repos (testing ColdIQ, Extruct, and the Claude GTM Plugin against real outbound workflows) are quite rigorous public evaluations of this stack.

Enterprise Account Executive at Centari (Hybrid - New York, NY)
Enterprise Account Executive, US & Canada at Semrush (Dallas, TX)
Senior Manager, Customer Enablement at CaptivateIQ (Remote - US)
Enterprise Account Executive at Cube (Hybrid - New York, NY)
Account Executive at Obvio (College Park, MD)
Founding Account Manager, Accountant at Kick (Toronto, Canada)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
RevStar Summit: June 3, 2026 (Toronto, Canada)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
CVC Week by Counterpart Ventures: September 29, 2026 (San Francisco, CA)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow Network love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Sophie Buonassisi and the GTMfund team (not AI!).

GTM: $2M ARR → $950M ARR → IPO. The sales story behind Figma.

GTMnow — Wed, 22 Apr 2026 16:04:44 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

If you spend time on X, you can also watch/listen to the entire episode on X.

Brought to you by:

One critical and surprisingly complicated decision is naming your startup. Founders often spend weeks chasing the perfect .com, only to overpay or settle for a name that doesn’t quite click.

Before you spend a ton securing a .com, we recommend checking out .tech domains.
.Tech tells the world - your customers, your investors, anyone Googling you: that you’re building in technology.

If you’re building a tech startup, secure your .tech domain today from any registrar of your choice. Learn more at https://startup.tech

Outbound sales today is still built around manual work: writing emails, building lists, and managing sequences across disconnected tools.

Nooks is breaking the mold. With their new AI Sequencing product, Nooks is an agent workspace where AI works alongside reps, helping them understand accounts, prioritize the right prospects, and generate context-rich outreach based on actual first-party interactions.

No more wasting time in multiple tools - it’s one unified workspace for all your outbound channels. Arm your sales team with intelligent outbound. Learn more at https://www.nooks.ai/gtmfund

Who we sat down with

Kyle Parrish joined Figma as the first-ever sales hire when the company was doing $2M in ARR and helped scale it to $950M ARR. Figma then went on to IPO (FIG).

But the path there was anything but smooth.

In this episode, Kyle breaks down what it actually took to build Figma’s enterprise sales motion from scratch, including the no-discount rule that made procurement teams furious, the 40-hour interview weeks when hiring felt impossible to keep up with, and what it was like to lead a 300-person team through a failed $20B Adobe acquisition, and then have their best year immediately after.

Discussed in this episode

Why Figma refused to discount, even when Microsoft pushed back
How to hire the right first sales rep as a founder
The PLG to enterprise transition most companies get wrong
What the Adobe deal collapse actually felt like from the inside
How Figma went from 3 products to 8 overnight and launched into an IPO
What great sales looks like in the AI era

Episode highlights

0:00 – Intro

1:02 – Guest intro: Kyle Parrish

1:35 – Joining Figma at $2M ARR in 2018

5:01 – First meeting with Dylan (Figma CEO)

6:13 – How to find your first sales hire

9:05 – Stage alignment in early hiring

11:43 – Early-stage operators need “scar tissue”

13:19 – Northstar metric at Figma

14:13 – Obsessing over customer conversations

16:44 – Building Figma’s brand through community

17:31 – Scaling the “unscalable”

20:08 – In-person GTM vs. digital

22:47 – Was there a moment Figma might not make it?

24:48 – Pivoting after the Adobe deal collapsed

28:10 – Figma’s no-discount rule

31:05 – Enterprise ELAs replacing discounting

34:47 – What makes a great salesperson in the AI era

36:56 – Missionaries vs. mercenaries in AI-era GTM

39:47 – Spotting the next Dropbox or Figma

43:46 – Kyle’s post-Figma investing focus

45:58 – Family, travel & what’s next

View the Full Transcript

Key takeaways

1. Stage alignment is everything when hiring your first sales rep.
Don’t bring in a big-company executive to an early-stage role. You need someone addicted to building, comfortable with ambiguity, and willing to do 15 jobs badly before figuring out which ones matter. The wrong hire at this stage can set your entire GTM culture back by years.

2. Figma’s no-discount rule was a trust play, not a pricing play.
By staying consistent across every customer, Figma built a reputation for transparency that became a competitive advantage. It was painful early, especially with enterprise procurement, but it forced the team to win on value instead of price.

3. PLG only works if your sales team respects the community it’s selling into.
Figma’s early sales motion was built on genuine curiosity, not quotas. Kyle and his team sat on couches with designers for hours, went to meetups, and learned the world they were selling into. That relationship with the community is what made the enterprise motion possible.

4. The Adobe acquisition collapse was Figma’s biggest unlock.
Being under a potential acquisition froze Figma’s ambition. The moment the deal died, the product team was unleashed, Dev Mode launched, and the company went from 3 products to 8 in a single conference. The $1B breakup fee funded the best chapter of Figma’s story.

5. In the AI era, sales fundamentals matter more, not less.
AI removes the excuse for poor preparation, but it puts an even sharper spotlight on how you build trust, do discovery, and show up in person. The sellers thriving right now are leaning into AI as a tool while doubling down on the human skills that no model can replace.

Follow Kyle Parrish

LinkedIn: https://www.linkedin.com/in/kparrish8
X (Twitter): https://x.com/KyleHParrish

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

It would mean the world to us if you subscribe to the show on your platform of choice. This helps support our growth, which in turn helps us continuously bring on bigger guests and deliver the content that you want to hear most.

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast shares how the best in tech build, scale and invest.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

GTMnow is run by GTMfund, an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Advice From the Founders of Slack, OpenAI, Stripe, and More.

GTMnow — Fri, 17 Apr 2026 21:29:10 GMT

This is a collection of reflections from founders who have collectively built over $3 trillion in enterprise value.

Each quote captures a defining moment: a decision, a near-breaking point, or a belief no one else saw. What follows is what they learned, and what you can apply.

Sales today still relies on manual work: stitching together “Frankenstacks” and toggling between a dozen tabs just to log a single call.

No more wasting time on data entry or disconnected tools. It’s one unified engine for the entire revenue cycle. Go Stackless. Learn more at reevo.ai.

Register here for the session - it’s time to orchestrate around how buyers actually buy.

1. Before anyone believes them

Every company on this list started as an idea most people dismissed. The common thread is not that these founders were contrarian for its own sake. They saw a specific problem with unusual clarity and refused to let consensus override conviction.

Jensen Huang · Co-founder & CEO, NVIDIA

Built NVIDIA from a graphics chip startup into a $3T+ computing platform powering the AI revolution.

“What is the problem you are trying to solve? What is the unmet need you believe will emerge? And what is it that you are going to do that is sufficiently hard that when everybody else finds out it is a good idea, they are not going to swarm it and make you obsolete?”

Huang said this, reflecting on NVIDIA’s founding in 1993, when the video game market was worth zero billion dollars. Sequoia’s Don Valentine told him startups should not invest in startups. He funded it anyway.

Melanie Perkins · Co-founder & CEO, Canva

Scaled Canva from a high school yearbook tool into a $40B design platform used by over 200 million people across 190 countries.

“Find the massive market hiding behind ‘too hard to use.’”

Perkins pitched to over 100 investors before raising her first round. The design software market was dominated by Adobe, and most investors could not see how a simplified tool could compete. She saw that the real market was not designers. It was everyone else.

Patrick Collison · Co-founder & CEO, Stripe

Built Stripe from a side project into a $95B payments infrastructure company processing over $1 trillion annually.

“For most products and most businesses, things can just be done much better. Moats are typically overrated.”

When the Collison brothers started Stripe in 2010, dozens of payment processors already existed. Everyone told them the market was solved. They saw that none of the existing products were actually good.

2. When the product finally clicks

Product-market fit is not a single moment, particularly now with AI (more on this coming soon). Despite the continuous nature of product-market fit now, there’s still usually a moment or inflection point. In the past, this would come once. Now, it comes far more frequently as products continuously adapt to evolving customer needs.

Stewart Butterfield · Co-founder & CEO, Slack

Pivoted from a failed video game into Slack, which grew to 750K+ paying organizations. Acquired by Salesforce for $27.7B.

“Sell the dream outcome. Know exactly what pain you are eliminating.”

Slack was born from the internal chat tool. Butterfield’s team built a video game called Glitch. The game failed. The chat tool became one of the fastest-growing enterprise products in history because it solved a pain point the team experienced firsthand.

Tobi Lütke · Co-founder & CEO, Shopify

Built Shopify from an online snowboard store into a $130B+ commerce platform powering over 10% of all U.S. e-commerce.

“The software I built for myself is good. Maybe other people want it too.”

Lütke wanted to sell snowboards online in 2004. Every e-commerce platform he found was, in his words, a user-hostile database editor. So he built his own. A year later, he realized the software was more valuable than the snowboard business.

Steve Jobs · Co-founder & CEO, Apple

Built Apple into the world’s most valuable company. Created the Mac, iPod, iPhone, and iPad.

“You cannot just ask customers what they want and then try to give that to them. By the time you get it built, they will want something new.”

Jobs said this in 1989, a decade before the iPod and two decades before the iPhone. His consistent thesis was that customers can describe their problems but not the solutions. The founder’s job is to close that gap.

3. Unlearning as needed

Scaling a company requires unlearning the very instincts that got you to product-market fit. The habits that work for 10 people often become liabilities at 100.

Tobi Lütke · Co-founder & CEO, Shopify

Has led Shopify through every stage, from two people in a coffee shop to over 10,000 employees.

“I did not believe in leadership. I just did not think leadership was important. I felt that if you get really good individuals, all the right things would happen. I believe those things are utterly incorrect now.”

Lütke admitted this on the Acquired podcast, reflecting on his early years as a programmer-turned-CEO. Scaling taught him that orchestrating great individuals into a team requires deliberate leadership, not just talent density.

Jensen Huang · Co-founder & CEO, NVIDIA

Retreated from the mobile phone market at its peak to bet on a then-zero-dollar AI-accelerated computing market.

“Retreating from a giant phone market to create a zero-billion-dollar market was risky, but it paid off. Strategic retreat is at the core of success.”

NVIDIA exited the mobile processor market around 2014 despite heavy investment. The decision freed resources that went into CUDA and GPU computing for AI. That strategic retreat is the foundation of NVIDIA’s current dominance.

Patrick Collison · Co-founder & CEO, Stripe

Scaled Stripe’s culture from 2 employees to 8,000+ while maintaining engineering-first principles.

“A rapidly scaling human organization is an unnatural thing. The cues and habits you learn from other human organizations are not going to be sufficient.”

Collison made this observation in an interview about culture at scale. His point: most human organizations (schools, churches, and families) do not double in size every year. Running a startup requires practices that feel contrived because the situation itself is unnatural.

4. The people who make or break it

Every founder on this list points to hiring decisions as the highest-leverage work they do. Not product. Not strategy. People.

Dharmesh Shah · Co-founder & CTO, HubSpot

Co-founded HubSpot and built the Culture Code deck, viewed over 5 million times. HubSpot now serves 200K+ customers.

“Build a movement first. The product follows the philosophy.”

Shah co-created the inbound marketing philosophy before the product fully existed. The movement attracted an audience of marketers who became the company’s first customers. Sometimes the community is the GTM strategy, and the product is what you build to serve it.

Sam Altman · CEO, OpenAI

Led Y Combinator as president, then built OpenAI into the company behind ChatGPT, reaching 400M+ weekly users.

“Mediocre engineers do not build great companies. You need to identify the people who are going to be really, really good, and go after them relentlessly.”

Altman repeated this principle across his YC lectures and Startup Playbook essays. His emphasis is on the disproportionate impact of the top 5% of talent, especially pre-product-market fit.

Marc Benioff · Founder & CEO, Salesforce

Grew Salesforce from four co-founders in a San Francisco apartment to over 70,000 employees and a $300B+ company.

“The secret to successful hiring is this: look for the people who want to change the world.”

Benioff’s hiring philosophy, outlined in Behind the Cloud, prioritized mission alignment over pedigree, which shaped the company’s early culture of customer obsession.

5. Drive growth at scale

Hypergrowth sounds like a dream until you are inside it. These founders describe the operational chaos, cultural strain, and identity crises that come with scaling faster than your infrastructure can keep up with.

Aaron Levie · Co-founder & CEO, Box

Built Box from a college dorm room into a publicly traded enterprise content platform with a $5B+ market cap.

“Compete on innovation, not features. Enterprise is won by whoever moves fastest.”

Levie started Box as a consumer product before pivoting to enterprise. The shift required rebuilding nearly everything. His takeaway was that speed of iteration matters more than feature count when competing against incumbents.

Parker Conrad · Co-founder & CEO, Rippling

After being removed as CEO of Zenefits, built Rippling into a $13.5B compound HR and IT platform from scratch.

“Compound products beat point solutions. Build the platform while others fight for features.”

Conrad’s thesis at Rippling is the opposite of conventional startup wisdom. He argues that the company that owns the system of record for employee data can build everything on top of it. Integration as moat.

6. When it feels impossible

There is a moment in every company’s history where the rational move is to stop. What separates these founders from the thousands who quit is not optimism. It is a specific kind of stubbornness backed by evidence that others could not see.

Jensen Huang · Co-founder & CEO, NVIDIA

Nearly went bankrupt in NVIDIA’s early years before the company found its path to dominance.

“I hope suffering happens to you. One of my great advantages is that I have very low expectations. People with very high expectations have very low resilience.”

Huang said this to Stanford students, explaining that NVIDIA nearly died multiple times in its first decade. Resilience is built by surviving difficulty, not by avoiding it.

Tobi Lütke · Co-founder & CEO, Shopify

Led Shopify through a near-meme-stock era during COVID, then restructured the company back to fundamentals.

“Ignorance is the single biggest driver of entrepreneurship. It was two years before it ever occurred to me that this might not work.”

Lütke says this without irony. His argument is that if you fully understood the odds, you would never start. Ignorance of the obstacles is a feature, not a bug, of the founding moment.

Demis Hassabis · Co-founder & CEO, Google DeepMind

Built DeepMind from a small London lab into the AI research organization behind AlphaFold, AlphaGo, and Gemini. Nobel Prize in Chemistry, 2024.

“I have always been interested in the biggest problems. If you are going to spend your time on something, it might as well be something that matters.”

Hassabis spent years building DeepMind before AI was commercially viable. When Google acquired the company in 2014 for over $500M, many questioned why a search company needed a fundamental AI research lab. The answer became obvious a decade later.

7. Words to remember on day 1

Hindsight compresses years of pain into clean sentences. These are the lessons these founders wish they had internalized from the start.

Jensen Huang · Co-founder & CEO, NVIDIA

33 years as CEO. Same company. Same mission. Different industry every decade.

“Strategy is not words. Strategy is action.”

Huang tracks what every employee’s top five priorities are and judges strategy by whether those priorities match the company’s direction. If the org’s daily actions do not reflect the strategy, the strategy does not exist.

Marc Benioff · Founder & CEO, Salesforce

Pioneer of the 1-1-1 philanthropic model, adopted by over 9,000 companies globally.

“The business of business is improving the state of the world.”

Benioff embedded the 1-1-1 model (1% of equity, 1% of product, 1% of employee time to philanthropy) into Salesforce from day one, before the company had revenue. Values are not something you add once you succeed. They are the architecture you build on.

Tobi Lütke · Co-founder & CEO, Shopify

Runs Shopify as a first-principles organization where every department must justify how it does things differently and better.

“The best thing founders can do is subtraction.”

Lütke’s advice to early-stage founders is counterintuitive: do less. The companies that win are not the ones that do the most things. They are the ones who identify the few things that matter and execute them with obsessive focus.

In a world where code writes itself and products can be copied overnight, the way you think is the last unfair advantage. The question is not whether you will face the same crossroads these founders faced. You will. The question is whether you will recognize it when you do.

Tag @GTMnow so we can see your takeaways and help amplify them.

AppDirect acquired PartnerStack, combining subscription commerce with partner and affiliate management in one platform. For companies running partner-led growth, it’s a meaningful consolidation — recruit, manage, and transact through partners without stitching together separate tools.

Anthropic hit a $3B revenue run rate and is doubling down on infrastructure, partnering with Google and Broadcom to build custom AI chips. The bet: owning silicon is how you control cost-per-token at scale, and at this growth rate, that math matters more every quarter.

VC: Your First VC Meeting Will Be Agent-to-Agent | Auren Hoffman (Flex Capital)

GTM 186: How One Hackathon Took Zapier’s AI Usage From 10% to 97% | CEO of Zapier

Listen by clicking the above links or by searhing wherever you get your podcasts “The GTMnow Podcast.”

Armada - named among Fast Company’s Most Innovative Companies in computing, alongside NVIDIA and Google. They’re building distributed AI infrastructure for the environments most platforms ignore. Worth watching as enterprises and governments push AI into the field, not just the cloud. To learn more you can also check out the GTMnow episode with CEO Dan Wright.

Examen - raised $4.3M in total funding and launched. Examen is building an autonomous analyst for commercial real estate.

Director, Marketing Operations at CaptivateIQ (Hybrid - Austin, TX | Remote - Raleigh, NC / Nashville, TN / Toronto, Canada)
Marketing & Advocacy Manager at Cube (Hybrid - New York, NY)
Enterprise Account Executive, US & Canada at Semrush (Dallas, TX)
Strategic Accounts at Coast (New York, NY)
Outbound Account Executive at Owner (Remote - US)
Enterprise Account Executive at Patch (Remote - San Francisco / New York)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
RevStar Summit: June 3, 2026 (Toronto, Canada)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
CVC Week by Counterpart Ventures: September 29, 2026 (San Francisco, CA)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow Network love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Sophie Buonassisi and the GTMfund team (not AI!).

GTM: 70% of the buying decision is made before they call you

GTMnow — Thu, 16 Apr 2026 15:47:26 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by:

Sales today still relies on manual work: stitching together “Frankenstacks” and toggling between a dozen tabs just to log a single call.

Reevo is breaking the mold. As a vertically integrated Revenue Operating System, Reevo replaces the fragmented maze with a single, AI-native workspace. It captures 100% of your first-party data, from emails to meeting recordings, to help reps find high-intent buyers, automate administrative “garbage,” and generate context-rich outreach that actually sounds human.

No more wasting time on data entry or disconnected tools. It’s one unified engine for the entire revenue cycle. Go Stackless. Learn more at reevo.ai.

Outbound sales today is still built around manual work: writing emails, building lists, and managing sequences across disconnected tools.

No more wasting time in multiple tools - it’s one unified workspace for all your outbound channels. Arm your sales team with intelligent outbound. Learn more at https://www.nooks.ai/gtmfund

Who we sat down with

AI isn’t just changing how we sell. It’s changing how buyers make decisions before they ever talk to you.

Sam Senior, Founder and CEO of TestBox, joins Sophie Buonassisi on The GTMnow Podcast to break down exactly what’s happening to the B2B software buying process right now, and what go-to-market leaders need to do about it immediately.

If you’re a founder, CRO, or AE wondering why your pipeline feels different, this conversation will give you a clear framework for what’s happening and what to do next.

Discussed in this episode

Why 70-80% of purchase decisions are already made before the first call (and it’s accelerating)
The shift from discovery calls to validation calls, and how to prepare
CEO (AI-version of SEO): how LLMs are shaping what buyers believe about your product
The “day one shortlist” shrinking from 3-4 vendors to 1-2 vendors
Agent-to-agent procurement: Sam’s timeline for when AI agents fully take over buying
Why the mid-funnel is actually getting longer, not shorter
The “Fake Nothing, Prove Everything” campaign that went viral post-Series A
How TestBox runs 15 AI experiments per week across the entire company
Using Google Vertex video analysis to read prospect body language on sales calls
How to build an AI-first culture without burning out your team
Books: The 15 Commitments of Conscious Leadership, No Ego, Courageous Marketing, Good to Great

Episode highlights

0:00 - Cold open

1:09 - What Test Box does

2:43 - How buying has changed

5:48 - What founders/CROs should do now

7:05 - GEO: AI version of SEO

9:02 - Why mid-funnel is expanding

21:17 - Agent-to-agent procurement

26:02 - All procurement by agents in 3-5 years

29:09 - How vendors differentiate beyond product

33:30 - The croissant campaign breakdown

45:50 - 15 AI experiments per week

47:10 - Analyzing prospects via video AI

48:29 - Building AI culture in your team

52:39 - Book recommendations

View the Full Transcript

Key takeaways

1. The first call is no longer a discovery call.
Buyers are doing so much research via LLMs before reaching out that by the time they talk to you, they’ve already made 70-80% of their decision. Your job on that first call is to validate (or correct) what they already believe, not discover their needs together.

2. If you’re not on the day one shortlist, you’ve already lost.
That shortlist has shrunk from 3-4 vendors to 1-2. The way to get on it is to audit what LLMs are saying about you right now. Ask Claude or ChatGPT what your buyers are researching and how your product shows up in those answers.

3. Agent-to-agent procurement is 3-5 years away, and the first steps are already happening.
Sam laid out a clear progression: LLM research today, AI-assisted trial evaluation in 12-18 months, agent-to-agent demos in 24-36 months, and full AI-led procurement including negotiation within 3-5 years. The companies building for this now will have a massive advantage.

4. The mid-funnel is getting longer, not shorter.
Buyers come in with high expectations because they’ve already TestBoxed your use case in ChatGPT or Claude and got 70-80% of the result in 30 seconds. Bridging that gap to 100% and proving your real value takes more time and trust-building than before.

5. AI adoption inside a company is a culture problem, not a tools problem.
Sam runs 15 AI experiments per week across the whole team, ties token usage to performance reviews, and runs a bi-weekly “Sneaks and Snacks” showcase. The shift only happened when he made it a company-wide operating model, not just an engineering initiative.

Follow Sam Senior

Linkedin: https://www.linkedin.com/in/samuelsenior
TestBox: https://www.TestBox.com

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast shares how the best in tech build, scale and invest.

GTMnow is run by GTMfund, an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

VC: Your First VC Meeting Will Be Agent-to-Agent

GTMnow — Wed, 15 Apr 2026 16:51:43 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by: AngelList

From starting as a small, operator-led rolling fund, to evolving to an institutional platform, AngelList has been a core partner in every phase of GTMfund’s growth. Their software-first fund admin infrastructure allowed us to scale without sacrificing agility — from onboarding hundreds of LPs seamlessly to handling compliance, capital calls, and reporting as our fund size evolved.

As we expanded from Fund I to Fund II, AngelList took care of the back-office operations, allowing us to stay focused on what matters most: investing in world-class founders and building the strongest go-to-market network in venture.

They’ve scaled with us across funds and into the future.

If your fund is growing in size or complexity, check them out at https://angellist.com/gtmfund.

Who we sat down with

Auren Hoffman (Flex Capital) joins the GTMnow podcast to share some of the most contrarian takes in tech today, from why AI moats are gone, to why your next VC meeting will be with a bot, to why AI is secretly going to trigger a baby boom.

Auren Hoffman is the founder of Flex Capital, SafeGraph, and LiveRamp. He's an early backer of Replit, Perplexity, Rippling, Vercel, Coinbase, Chime, and AppLovin.

Discussed in this episode

Why Auren runs 500+ AI agents to source deals, and what that means for founders raising capital
The “agent-to-agent” meeting prediction: by end of 2026, first VC conversations will be fully automated
Why every software moat has been “blown up” and what Salesforce, LinkedIn & DocuSign need to do to survive
The OpenAI x The Hustle acquisition breakdown: why it’s the smartest (and cheapest) distribution play in AI
Why missing a great deal is 10x more painful than making a bad one, Auren’s honest VC mistake framework
The baby boom thesis: why AI, IVF, self-driving cars & cheaper energy could reverse the fertility decline
Why companies won’t sign yearly SaaS contracts anymore, and what that means for every B2B founder

Episode highlights

01:05 - GTMfund Q1 recap

02:38 - OpenAI x The Hustle breakdown

06:18 - Redpoint’s optimal VC deployment period

11:24 - Auren Hoffman intro

13:04 - Why am I seeing this deal?

26:26 - Sizing up founders at Replit, Perplexity & Rippling

28:49 - What separates great founders

32:10 - 500+ AI agents for deal sourcing

33:40 - Agent-to-agent VC meetings by 2026

45:13 - Every software moat is blown up

49:09 - Who kills Salesforce next?

51:30 - Why no one signs yearly SaaS contracts anymore

51:50 - AI will trigger a baby boom

56:22 - Thinking generationally

View the Full Transcript

Key takeaways

1. Every software moat is gone.
If you’re not making your product significantly better every month, you will lose customers. The era of “good enough” SaaS is over: Salesforce, LinkedIn, DocuSign are all vulnerable.

2. Agent-to-agent meetings are coming by end of 2026.
Flex Capital already uses 500+ AI agents for sourcing. Auren predicts the first VC meeting will soon be fully automated, founders and investors talking through their agents before any human interaction happens.

3. Missing a great deal is 10x worse than making a bad one.
The biggest VC mistake isn’t backing a loser — it’s never seeing the winner in the first place. Auren’s whole system is built around seeing as many companies as possible, not filtering too early.

4. Always ask “why am I seeing this deal?”
The best companies don’t need you. If a founder is letting you write 80% of their seed round, that’s a red flag, not an opportunity. Deal flow quality is everything.

5. AI will trigger a baby boom among wealthy Americans.
Counterintuitive but data-backed: married, higher-income Americans are already having more kids. As AI reduces the cost and complexity of child-rearing, through robotics, self-driving cars, and cheaper energy, Auren believes that trend accelerates dramatically.

Follow GTMfund

Newsletter: You’re in it :) If this was forwarded to you, you can subscribe here. GTMnow is our main brand for content, but we also have two around our investments:
LinkedIn: https://www.linkedin.com/company/gtmfund
X: https://x.com/gtmfund

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast shares how the best in tech build, scale and invest.

GTMnow is run by GTMfund, an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

AI Conference HumanX Takeaways and Frameworks

GTMnow — Fri, 10 Apr 2026 20:20:01 GMT

6,714 AI leaders from around the world just gathered for a massive AI conference in San Francisco. There were people from 79 countries and 500+ sessions.

From over 35,000 minutes of content at the event, these are the 9 takeaways to know. There are also frameworks and mental models after the takeaways - one you absolutely need to know is the three waves of AI by Jensen Huang.

Every sales leader knows the story: your best reps carry the number while everyone else tries to get lucky. HockeyStack closes the gap with Revenue Agents for the enterprise.

Every deal gets its own dedicated agent, surfacing the right play based on how your company actually wins. Reps focus on judgment, relationships, and the conversations that close deals. Agents handle everything else.

That’s why teams like RingCentral, Outreach, ActiveCampaign, and Fortune 100 companies rely on HockeyStack. Learn more at hockeystack.com

Outbound sales today is still built around manual work: writing emails, building lists, and managing sequences across disconnected tools.

Nooks is breaking the mold. With their new AI Sequencing product, Nooks is an agent workspace where AI works alongside reps, helping them understand accounts, prioritize the right prospects, and generate context-rich outreach based on actual first-party interactions.

No more wasting time in multiple tools - it’s one unified workspace for all your outbound channels. Arm your sales team with intelligent outbound. Learn more at https://www.nooks.ai/gtmfund

Takeaways

1. Spatial intelligence is the next frontier.

Dr. Fei-Fei Li (World Labs founder, Stanford AI Institute) delivered a thesis that will define the next five years of AI investment: everything we’ve built so far is “wordsmiths in the dark.” Language models can generate text brilliantly. But they have zero understanding of the 3D world, physics, movement, or causality. They’re entirely abstract.

Spatial intelligence – the ability to perceive, reason, and generate in actual 3D space – is the missing layer separating AI that can do real work from AI that can only theorize about work. A language model can’t teach a robot to move, a text generator can’t build a dynamic game world, and an abstract medical algorithm can’t diagnose from 3D scans without spatial reasoning. You can’t build intelligence on language alone.

This creates a data constraint. Spatial training data is exponentially scarcer than language data. Which means the companies that crack the synthetic data flywheel first – building models that generate realistic 3D worlds that then train robots and other models, creating new training data – will own the next decade the way foundation models owned this one.

If your application doesn’t interact with the physical world, language models are probably sufficient. If it does (robotics, supply chain, manufacturing, design, healthcare), you need spatial grounding. You need partners building spatial intelligence. The vendors dominating spatial intelligence in the next 5 years will be the infrastructure layer of the 2030s.

2. There are 3 waves: Generative → Reasoning → Agentic.

Jensen Huang (NVIDIA) mapped the actual trajectory of AI that most people misunderstand.

Wave 1 (2023-2024) was generative: can you make text, images, code?
Wave 2 (2024-2025) is reasoning: can you think through problems, condition outputs, do research?
Wave 3 (2025-2026+) is agentic: can the model take a business problem in plain English and execute end-to-end across your actual systems?

The pattern is in the prompts.

Year 1: “What is?” “Where is?” “How does?” (extraction).
Year 2: “Reason about this.” “Summarize this.” (thinking).
Year 3: “Build this for me.” “Execute this.” “Figure it out and report back.” (autonomous work).

His prediction flips the org chart: within 1-2 years, a non-technical CEO will be able to describe a business problem and have an agent execute the solution. Not draft an email – actually integrate systems, read documentation, iterate, solve. You don’t need engineers to orchestrate anymore, you need a person who understands the outcome and can articulate it to the agent.

This collapses the entire skill ladder. Traditionally, engineers have been the bottleneck. They connect systems, pull data, build workflows. In 12-18 months, someone describes the outcome and agents build the workflow. This is a categorical shift.

You’re shouldn’t be waiting for agents to be “ready,” because they’re already ready for 70% of your workflows. You’re betting on orchestration and adoption, not technology. The teams that move fastest now – that start orchestrating agents into their core workflows in the next 3 months – will set the pattern for how their entire industry operates. Everyone else will be copying.

3. The skill becoming less valuable is execution, and the skill becoming valuable is taste.

Inside OpenAI, Srinivas Narayanan’s engineers don’t write code anymore. They guide agents that write code. 80% percent of their time is spent on judgment – knowing what the AI did wrong, what to try next, whether the output is accurate. A full repository is generated by CodeX, but the human decides if it’s right.

Open Evidence’s 70 engineers built what would have taken 500 because the execution layer collapsed.

What’s left? Taste.

This inverts your hiring and promotion calculus. You can’t hire for “fast executor” anymore. That’s a commodity. You need people with strong opinions about the domain, the ability to see flaws in AI output before they ship, and judgment at speed.

If you’re still optimizing for speed to execution, you’re hiring for yesterday’s bottleneck.

4. “Human in the lead, not in the loop” is how enterprises scale.

When enterprises say “human oversight,” they usually mean “humans constantly validating every decision.” That doesn’t scale. The competitive advantage is human in the lead. This is where humans set vision, strategy, what to measure. Then, agents handle execution at scale.

This can impact who you’re selling to and how. If you were targeting IT or operations heads who look to control variables, you may now be pitching business leaders who want to set direction and move at venture speed. Different buyer, different message, completely different sales cycle. Be mindful of this distinction.

5. The model is free now. What’s actually scarce is problem insight.

Open Evidence shouldn’t exist. 70 employees, 35 engineers, $12+ billion valuation, 60%+ of US physicians using it daily. If owning the model was the defensible asset, a startup with no custom models should be impossible. But they don’t own the model, they own the problem.

The entire narrative around “model superiority” has inverted. Mark Terbeek (Greycroft) and Hans Tung (Notable Capital) see the same pattern across winners: the companies winning aren’t the ones with the best technology, they’re the ones closest to the customer problem. The ones that built backward from “what do doctors actually need” instead of “what can our model do?”

Speed kills everything else now. When the model improves weekly, shipping fast and staying close to customer problems beats feature parity every single time. A smaller team moving at venture speed will always outrun an enterprise that needs to deploy changes through committee.

This kills the traditional moat-building playbook. You can’t win on technology ownership. Models are commodities improving on a one-week cycle. What’s defensible is problem insight – the founders and team who understand the domain so deeply that they know what to build before customers can articulate it. And the organizational speed to ship it before the model catches up and makes it free.

6. Unit economics are what separate durable companies from trends.

Salil Deshpande (solo GP managing $750M) named the crisis everyone’s avoiding: most AI infrastructure has negative gross margins. Models improve weekly, closed-source models stay 9 months ahead. The math doesn’t work. This is dot-com infrastructure all over again – massive overbuilding followed by a collapse that wipes out companies built on trends, not unit economics.

When you can’t sustain your CAC, improve margins every quarter, or scale profitably, you’re not a business. You’re a feature waiting to be copied or a vendor waiting to disappear.

For operators evaluating vendors: if a vendor has sub-70% gross margin, they’re not thinking about durability. They’re thinking about growth at any cost. Which means within 18 months they could pivot, get acquired, shut down, or be forced to raise at a down round and cut support. You’ll be stuck with a tool that became orphaned. Ask about CAC payback and gross margin trajectory. If they hedge or refuse to answer, that’s your signal.

7. Agents don’t replace your team, they replace your processes.

Madhav Thattai (Salesforce, Agentforce) and Rob Seaman (Slack) exposed a pattern that kills the “AI will displace workers” narrative: Agentforce is at $800M ARR with 25,000 customers running billions of agent transactions. Companies aren’t shrinking, they’re moving faster and becoming more ambitious.

The fear is misplaced. Agents don’t replace your team, they expose what your team should actually be doing. When an agent handles the 80% of customer service that’s rote, the remaining 20% becomes visible: the cases requiring judgment, empathy, problem-solving. That human? They’re worth more now, not less. Engineers don’t disappear when agents write code. They become the people who brief agents on what to build next, review architecture, make strategic calls.

8. Where agents sit matters more than what they do.

Slack hit a billion messages a day. 1,000% increase in AI apps being built. And the agents didn’t get smarter, they just moved. That’s the entire story!

The gap between an agent people use and an agent people ignore is placement. An agent on a separate website? Adoption dies. That same agent, invisible in Slack where work actually happens? Adoption jumps 25% immediately.

This is where most AI investments fail: they’re building brilliant agents and then burying them three clicks away in a separate tool. Or worse, they’re expecting sales teams to “discover” the agent, log in to a new system, learn a new interface. That’s never going to compete with the agent that just appears in Slack when someone says “I don’t know what to do.”

Rob Seaman explained how Slackbot becomes the invisible router: when someone needs benefits info, Slackbot surfaces the benefits agent. When they need to file a ticket, it routes to Linear. The agent never needs to be named or discovered, it just emerges contextually from the flow of work.

For product builders: your agent’s capability is probably fine. Maybe it’s 70/10/20. But if it’s not in the flow of work – if it requires an extra step to access – you’re betting against human friction. Humans are lazy. They use what’s in front of them. Build the agent, then architect the placement.

9. Day 2 is harder than day 1…observability is the real battle.

Every founder thinks Day 1 is the hard part: build the agent, get it live, ship it. That’s when the real work starts, and most teams aren’t prepared for it.

Companies sprint to build an agent in 2-3 weeks, celebrate the launch, and then discover they have no visibility into whether it’s actually working. The agent is live. Is it performing? Is it handling the right cases? Is it drifting? In healthcare and financial services, “mostly works” is code for “about to create a lawsuit.”

The difference between a high-growth company and a stalled one is observability. Most vendors obsess over agent capability. The ones winning obsess over measurement. Companies need to track agentic work units (actual completed tasks, not tokens), monitor KPI delivery, and spot drift the moment business conditions change but the agent still follows instructions from three months ago.

Salesforce itself learned this the hard way: building the agent took 1.5 months. Actually operating it – refining, measuring, optimizing – took another 2+ months. And it’s continuous. When a KPI changes, the agent gets coached and re-measured. When business rules shift, the agent doesn’t auto-adapt. Someone has to notice and intervene.

The vendors pitching “build agents in hours” are selling you the first mile. The vendors talking about observability, monitoring, and Day 2 operations are selling you the other 99.

Frameworks & Mental Models

These are the reusable patterns underneath the takeaways. Use these to think about your own situation.

The Five-Layer AI Stack (Jensen Huang, NVIDIA)

Power → Chips → Infrastructure → Models → Applications

Every layer has its own ecosystem, margins, and competitive dynamics. Most venture capital flows to applications, but applications are worthless without the layer below them. The constraint shifts over time. Right now, power and chip design are bottlenecks. In 18 months, it might be models. In 3 years, it might be applications.

A critical insight: the most important layer is applications. Not because it’s the sexiest, but because it’s where value accrues for customers. Chips are commodities unless they enable new applications. Models are infrastructure unless they unlock new work. Energy is useless unless it powers something people want.

When evaluating a company or a vendor, ask: which layer are they actually winning in? Are they defending a layer they’re good at or pretending to compete everywhere?

The Three Waves of AI (Jensen Huang)

Wave 1: Generative (2023-2024)

What it does: Generate text, images, code from language
Prompt pattern: “What is?” “Where is?” “How does?”
User behavior: I ask, AI answers
Business impact: New content, new marketing, new coding velocity

Wave 2: Reasoning (2024-2025)

What it does: Reason through problems, do research, condition outputs
Prompt pattern: “Summarize this.” “Reason about this.” “Why did X happen?”
User behavior: I ask complex questions, AI reasons aloud
Business impact: Better accuracy, grounded outputs, trust increases

Wave 3: Agentic (2025-2026+)

What it does: Take business problems in natural language and execute end-to-end
Prompt pattern: “Create X for me.” “Build Y.” “Execute this task and report.”
User behavior: I describe outcome, AI does the work autonomously
Business impact: Discontinuous productivity, new categories of work, skill compression

If you’re still operating in Wave 1 or Wave 2 thinking, you need to get to Wave 3.

Spatial Intelligence vs. Language Intelligence (Fei-Fei Li)

Language Intelligence (today)

What it understands: Text, patterns in text, how to generate new text
What it’s blind to: 3D space, physics, movement, causality in the physical world
Metaphor: “Wordsmiths in the dark”—brilliant at language but not grounded in reality

Spatial Intelligence (emerging)

What it understands: 3D space, geometry, physics, dynamics, movement, interaction
What it enables: Robotics, autonomous systems, medical imaging, game worlds, design tools
Output: Not text or images, but 3D worlds, simulations, predictions of next states

The convergence: Language models + spatial models together = AI that can reason about work and execute it in the physical world. Language alone is theory. Spatial alone is mechanics. Together, they’re intelligence.

The Maturity Curve: Automation → Discovery → Real Work (Salesforce/Slack)

Every agent deployment follows this progression:

Month 1: Task Automation

Agent handles simple, repetitive tasks
Status checks, email drafting, policy lookup
User sees: “The system can do this for me”
Business value: Time savings on low-impact work

Month 2: Information Discovery

Agent can answer questions, synthesize information, find context
Customer history, policy explanations, data lookups
User sees: “The system understands my domain”
Business value: Faster decision-making, less context-switching

Month 3+: Real Work

Agent executes consequential tasks: orders, service resolutions, transactions
Full workflow autonomy with guardrails
User sees: “The system actually moves business forward”
Business value: Revenue impact, customer experience transformation

Don’t judge agents on Month 1 capability. They all look mediocre. The companies winning are the ones that have the infrastructure and discipline to ship Month 3 work.

Agentic Work Units vs. Tokens (Salesforce)

Tokens = input mechanism

How many words/pieces of text does the model consume?
Like measuring a truck by how much fuel it burns, not how much it delivers

Agentic Work Units = output mechanism

How many meaningful tasks did the agent complete?
Actual business outcomes: orders processed, customers resolved, decisions made

The mistake: optimizing for token efficiency instead of work unit impact. You can burn tokens forever without delivering value.

Three Constraints on Frontier Models (Fei-Fei Li)

For any frontier AI company, three things constrain growth:

Compute: Access to GPUs, data centers, training infrastructure
Models: Research capability to innovate new architectures
Data: Training data that teaches the model what actually matters

Most companies obsess over compute and models. Data is the hidden constraint. Spatial data is even scarcer than language data. The data flywheel—where output becomes input for the next generation—is where defensibility lives.

Two Ways to Think About “Busy” (Jensen Huang)

Prescriptive: “Here’s exactly what you should do with AI”

Pros: Clear, measurable, predictable
Cons: Misses most of the opportunity, forces rigid thinking

Inspiring: “Here’s why this matters. I can’t predict exactly how you’ll use it, but this is the direction.”

Pros: Captures optionality, lets teams find their own breakthroughs
Cons: Harder to measure, requires trust in distributed judgment

The best leaders let a thousand flowers bloom everywhere except the critical core (coding, supply chain, chip design). In those domains, you go to the frontier and don’t fail.

HumanX

Following a landmark second edition at the Moscone Center in San Francisco, HumanX 2027 will take place from March 7–10, 2027, at Mandalay Bay in Las Vegas.

Tag @GTMnow so we can see your takeaways and help amplify them.

Trimble is acquiring Document Crunch, bringing AI-powered contract risk analysis into its construction project management ecosystem. Construction is one of the last industries to modernize around document compliance, and purpose-built AI that catches payment disputes and notification failures before they escalate is a real wedge. Vertical AI finding its way into incumbent platforms is the pattern to watch.

GTM 185: How One Hackathon Took Zapier’s AI Usage From 10% to 97% | CEO of Zapier

For the full thing, listen on Apple, Spotify, YouTube or wherever you get your podcasts by searching “The GTMnow Podcast.”

Examen - raised $4.3M in total funding and launched. Examen is building an autonomous analyst for commercial real estate.

Enterprise Account Executive, US & Canada at Semrush (Dallas, TX)
Enterprise Account Executive at Patch (Remote - San Francisco / New York)
Enterprise Customer Success Manager at Noibu (Hybrid - Ottawa, ON)
Growth Strategist at Mutiny (New York, NY)
Senior Account Executive at Statusphere (Remote - Orlando, FL)
Senior Account Executive, Brands at Fastbreak AI (Charlotte, NC)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

MicroConf 2026: April 12–14, 2026 (Portland, OR)
GTMfund Dinner: April 14, 2026 (Austin, TX)
HockeyStack’s Launch Party: April 16, 2026 (San Francisco, CA)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
CVC Week by Counterpart Ventures: September 29, 2026 (San Francisco, CA)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow Network love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Sophie Buonassisi and the GTMfund team (not AI!).

GTM: How Zapier got to 97% company-wide AI adoption

GTMnow — Wed, 08 Apr 2026 16:58:59 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by:

Every sales leader knows the story: your best reps carry the number while everyone else tries to get lucky.

HockeyStack closes the gap with Revenue Agents for the enterprise. Every deal gets its own dedicated agent, surfacing the right play based on how your company actually wins. Reps focus on judgment, relationships, and the conversations that close deals. Agents handle everything else. That's why teams like RingCentral, Outreach, ActiveCampaign, and Fortune 100 companies rely on HockeyStack.

Learn more at hockeystack.com.

Who we sat down with

Wade Foster is the CEO of Zapier, a company that sits between 7,000+ apps and runs millions of automations every single day. That gives him a front-row seat to how companies are actually adopting AI, not just talking about it.

In this episode, Wade breaks down the exact decisions he made at Zapier to go from 10% AI usage to 97% company-wide, why agents and workflows are not the same thing, and what most leaders are getting completely wrong about AI fluency.

Discussed in this episode

The difference between agents and workflows (and when to use which)
What triggered Zapier’s internal “Code Red” after GPT-4 launched
The one-week hackathon that took AI adoption from 10% to 50% overnight
The AI fluency rubric Zapier built: Unacceptable, Acceptable, Adaptive, Transformative
Why leaders who aren’t using AI are the biggest bottleneck in their companies
How to measure AI ROI: floor raisers vs ceiling raisers
How AI now handles 50% of Zapier’s customer support tickets
Wade’s personal “advisory council” of AI sub-agents he uses for every major decision
Why building a company today is 10x cheaper but distribution is 10x harder
The truth about fundraising: you’re selling your company, not raising money
How Zapier stayed profitable by only hiring when it hurt

Episode highlights

0:00 - Intro

1:13 - The Seinfeld Quote & Kanye Text story

3:16 - Workflows vs. Agents: What’s the difference?

6:09 - Zapier’s Code Red moment

8:55 - The hackathon that moved AI adoption from 10% to 50%

12:06 - Making AI fluency a hiring requirement

13:47 - Building the AI fluency rubric

16:40 - Why leaders are the biggest AI bottleneck

18:12 - Revenue impact of going AI-first

22:09 - Would Wade build Zapier differently today?

23:20 - Is Zapier’s moat at risk from agents?

24:59 - Staying profitable with minimal capital

28:37 - The riskiest contrarian bet that paid off

31:52 - Wade’s 3 personal AI workflows

36:53 - Favorite books for founders

View the Full Transcript

Key takeaways

1. Agents and workflows are not the same thing, and confusing them is costing companies.
Workflows are deterministic, they behave the same way every time. Agents get a goal and figure out how to complete it on their own. Most companies are reaching for agents when a simple workflow would be cheaper, faster, and more reliable. Know which tool the job actually needs.

2. The fastest way to drive AI adoption company-wide is a hackathon, not a mandate.
Zapier went from under 10% to over 50% daily AI usage in a single week by running a company-wide hackathon across every function, not just engineering. Telling people to use AI does nothing. Giving them a dedicated week to play with it changes behavior permanently.

3. If your leaders aren’t using AI themselves, they can’t lead the transition.
Wade’s sharpest point: the reason most companies struggle to build an AI fluency rubric is that their leaders have never used the tools. You can’t define “good” if you’ve never done it. Leaders who are still running a 2019 playbook are the single biggest bottleneck to company-wide adoption.

4. Stop measuring AI productivity. Measure business outcomes instead.
AI ROI has two layers: floor raisers (everyone using AI daily, hard to quantify but real) and ceiling raisers (specific, measurable wins like 50% of support tickets resolved by AI with higher CSAT). Track your actual KPIs and work backwards. AI for the sake of AI is just expensive experimentation with no accountability.

5. Building is 10x cheaper today. Distribution is 10x harder. That changes everything. The cost to build a product has collapsed with AI. But getting anyone’s attention in a saturated market is harder than it has ever been. If you’re starting a company in 2026, your product advantage disappears fast. Your distribution strategy is the actual moat.

Follow Wade Foster

LinkedIn: https://www.linkedin.com/in/wadefoster
Zapier’s LinkedIn: https://www.linkedin.com/company/zapier
Zapier’s website: https://zapier.com

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast shares how the best in tech build, scale and invest.

GTMnow is run by GTMfund, an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

The 0 to 1 Guide to Paid Media

GTMnow — Fri, 03 Apr 2026 17:43:48 GMT

The news yesterday that OpenAI acquired TBPN truly demonstrates how valuable distribution is. As it’s increasingly more valuable but also more challenging to build, many companies and leaders have been inquiring more about paid ads. So, we teamed up Rex Gelb, Founder at Summit Chase and Head of Paid Media at Cursor, to bring you this edition.

Rex spent over a decade leading paid media at HubSpot. He started as a team of one managing a single Chrome extension on a $20k/month budget. By the time he left, he had built and led a global team of 20+ operators running campaigns across dozens of products, 15+ countries, and 6 languages.

Over 12+ years, he has managed roughly $750M in spend across Google, Meta, LinkedIn, and other major platforms. Now, he’s leading performance marketing at Cursor.

This guide walks through the 0 to 1 playbook. When to start, how much to spend, what to test first, and how to scale once something finally works. Let us get into it.

Sales today still relies on manual work: stitching together “Frankenstacks” and toggling between a dozen tabs just to log a single call.

No more wasting time on data entry or disconnected tools. It’s one unified engine for the entire revenue cycle. Go Stackless. Learn more at reevo.ai.

Step 1: Know When Not to Run Ads

The biggest mistake founders make with paid media is starting too early.

Ads are not the tool that creates product-market fit. They amplify it.

If you haven’t reached basic traction yet, paid media usually just accelerates the learning that something isn’t working. Rex has seen this pattern hundreds of times: founders pour $30K-50K into ads before retention exists, then conclude paid media does not work for them. The reality is the product was not ready. The ads just made that visible faster. Without product-market fit, you are paying to acquire users who churn immediately, and the platform learns to find more users who look like your churning users. It becomes a negative feedback loop.

If retention is not there yet, you are probably too early. Paid media works best when it pours gasoline on a fire that already exists.

Before investing in ads, you should already see signs that the product works:

Users are sticking around. Not just signing up, but coming back. If your week-4 retention is below 20%, ads will amplify a leaky bucket.
Customers are renewing. If customers are not staying, spending more to acquire new ones is a losing formula.
Word of mouth exists. People recommending your product without being asked.
Early deals are happening organically. Revenue coming in through inbound, referrals, or outreach that converts without heavy discounting. If you can’t close deals at all, paid media will not fix your sales motion.

Paid media works best when it pours gasoline on a fire that already exists.

The exception is early product testing. Some startups will run small paid campaigns to quickly recruit beta users or gather feedback. In that case the goal isn’t revenue - it’s learning. Small campaigns ($2K-5K) to recruit beta users, test messaging, or validate positioning are fine. But large-scale pipeline generation should wait until traction exists.

Step 2: When You Have Traction, It’s Almost Always Worth Testing

Once you have product-market fit and some capital to deploy, paid media becomes a very asymmetric bet.

The downside of testing is relatively small. You spend $20K-50K over a few months and learn what does not work. The upside is a repeatable pipeline engine that compounds for years. That risk-reward profile is why most Series A companies should be testing paid.

Most B2B companies start testing paid media around Series A, when they have both traction and budget.

Typical starting ranges look something like this:

Seed / Pre-PMF ($0-20K/month)
Mostly small tests and signal gathering. At this stage, paid media is a research tool, not a growth engine. Rex started at this level at HubSpot: one person, one product, $20K/month budget. The goal was to figure out whether paid could work at all.

Series A ($20-50K/month, ramping to $100K+)
Aggressive teams may ramp toward $100K+ if tests work. This is the inflection point. Test 3-4 campaign types simultaneously; measure pipeline, not just leads. Two to three months is typical before something works consistently. Resist the urge to quit after four weeks.

Series B+ (often $200K/month and up)
This is common depending on growth targets. At this stage you are scaling what works: new channels, geos, audience segments. Key: CPC has risen for 87% of industries (WordStream). If a channel produces ROI today, scale aggressively now, not next quarter.

These are not rules, they are just patterns.

Step 3: Expect It Not to Work at First

The first few months of paid media are usually messy. Rex shared at HubSpot’s INBOUND conference that roughly 90% of Facebook ad campaigns fail when they launch. The bar is not “get it right the first time.” Rather, it’s to “build a system that finds what works faster than competitors do.”

The goal early is not efficiency. The goal is finding signal.

Campaigns miss targets, messaging feels off, targeting is wrong. This is normal. Remember, the goal is finding signal and learning.

Winning ad programs come from testing combinations across five variables. Think of it as a five-dimensional matrix. Each combination you test narrows the space of what works.

The five variables to test:

Platforms. Which channels does your audience actually respond on? LinkedIn might seem obvious for B2B, but Rex has seen Google competitor campaigns outperform LinkedIn for certain products. Do not assume. Test.
Audiences. Job titles, seniority levels, company sizes, industries. Rex consistently finds that the buyer you think will convert best is not always the one who actually does.
Creative. Long copy vs. short copy. Video vs. static. In 2026, authentic native-style content increasingly outperforms polished production, especially on Meta.
Offers. Demo request vs. free trial vs. gated content. A demo request typically converts to pipeline at 3-5x the rate of a whitepaper download.
Landing pages. On LinkedIn specifically, Lead Gen Forms increase completion rates dramatically by auto-filling user data.

Eventually something starts to click. Then you double down.

The Core Three: Platform-by-Platform Playbook

Below is a practical starting framework for the three platforms most B2B companies should test first. These cover the three primary buying modalities: professional targeting (LinkedIn), active search intent (Google), and broad reach at scale (Meta). Most B2B companies should master these three before expanding to YouTube, Reddit, Twitter/X, or CTV.

LinkedIn Ads

LinkedIn is usually the first paid channel B2B teams test. It offers the cleanest targeting for professional audiences. You can target by job title, seniority, company size, industry, and even specific companies. No other platform comes close to this level of professional precision.

But the biggest mistake companies make on LinkedIn is adding unnecessary friction. Driving traffic to a landing page sounds logical, but it often kills conversion rates. The user has to leave LinkedIn, wait for the page to load, manually fill out a form, and submit. Each step loses people.

Instead, start with LinkedIn Lead Gen Forms. These keep the user inside LinkedIn and auto-fill their information (name, email, company, title) from their LinkedIn profile, which dramatically increases completion rates.

What to test on LinkedIn:

Thought leadership ads. Short insights from founders, operators, or subject matter experts consistently outperform polished marketing copy. LinkedIn now supports Thought Leader Ads that run from personal profiles rather than company pages. Early data shows CPCs of $5-$15 with meaningfully higher engagement rates.

Document Ads. These are an underused format that deserves more attention. You upload a PDF (whitepaper, guide, data report) and users can preview it directly in their feed without leaving LinkedIn. The download interaction pre-qualifies the lead. Document Ads deliver the lowest average CPL on the platform: $256, compared to $317 for single image ads.

In-feed ads. Simple posts that appear naturally in the feed. These work well for demo requests, product announcements, and event promotions.

Message or conversation ads. Direct outreach style ads that mimic a sales conversation.

Content offers or demo requests. Test both gated content and demo CTAs.

Also test multiple personas. sometimes the buyer you think converts best is not the one who actually does. Run the same offer to 3-4 different persona segments and let the data tell you who your real buyer is. You might discover that mid-level managers convert better than VPs, or that a specific industry segment responds at 3x the rate of your assumed ICP.

LinkedIn success comes from systematic testing. 2026 benchmarks: Avg B2B CPC $5.50-$8.50. CTR 0.44-0.65%. B2B ROAS 4.1-8.3x. LinkedIn now captures ~39% of all B2B paid media budgets globally.

Google Ads

Google is often the highest-intent paid channel. But many companies launch Google campaigns incorrectly. Always start with Search ads, not display or YouTube.

There are three core campaign types to test:

1. Brand Campaigns

If competitors are bidding on your company name, they can capture demand you created. Brand campaigns protect that traffic.

It may feel strange paying for clicks on your own name. After all, your organic listing is right there. But in practice, competitor ads push your organic result further down the page, and even a small percentage of clicks going to competitors represents real pipeline leakage. Brand campaigns are usually very cheap (low CPC, high CTR, high conversion rate) because the user already has intent to find you.

2. Competitor Campaigns

Competitor terms are often one of the highest ROI tests in early Google programs.

The conversion rates on competitor campaigns tend to be lower than brand campaigns (they are looking for someone else, after all), but the lead quality is often exceptional because these are buyers deep in an active evaluation. The key is having a compelling reason for them to consider you: a clear differentiator, a comparison page, or a case study from a customer who switched.

3. Non-Brand Search

These are category keywords like “HR compliance software,” “B2B lead generation platform,” or “marketing attribution tool.”

These can be competitive and expensive, but they help you capture new demand. Other Google formats like Performance Max, Display, and YouTube can work well later. But for 0 to 1, Search should always come first.

Meta (Facebook and Instagram)

Meta ads behave very differently than LinkedIn and Google.

The platform relies (even more) heavily on machine learning. The ad system learns from massive volumes of behavioral data and makes targeting decisions that often outperform manual audience selections. Trying to micromanage campaigns or over-segment audiences usually makes performance worse, not better.

Instead, take a consolidated approach. This means running a small number of campaigns with broad targeting and letting the algorithm optimize. Specifically:

Do not separate Facebook from Instagram. Let Meta decide where your ad performs best for each user. The platform has more data about user behavior across its properties than you do.
Do not split mobile from desktop. Again, the algorithm handles this better than manual segmentation.
Do not create tiny audience slices. A campaign targeting 5,000 people gives the algorithm nothing to work with. Broader targeting (50K-500K+) lets the machine learning system find your buyers within a larger pool.
Avoid using lead forms unless you include strong qualification questions. Meta lead forms often produce high volume but low quality because the auto-fill makes it too easy. Driving traffic to your website with a clear conversion action (demo request form, free trial signup) usually produces better-qualified leads.

The 2026 creative shift: Meta’s algorithm increasingly favors authentic, native-style content over polished production. Short-form video with CPVs as low as $0.01-$0.02 is becoming the default. A founder recording a quick take on their phone may outperform a two-week agency production.

Always measure pipeline contribution, not just top-of-funnel volume.

After the Core Three, Expand Carefully

Once you have data from LinkedIn, Google, and Meta, additional channels may be worth testing.

Common next experiments include:

YouTube for video education and retargeting
Reddit for niche technical communities
Twitter/X for startup and developer audiences
CTV for brand awareness at scale

But most B2B companies should focus on the core three first.

Spreading budget across too many channels early slows learning. The math: $30K/month across three platforms gives you $10K per platform. $30K across six gives you $5K each – often below the threshold for statistically meaningful results on any of them.

The Most Important Step: Set Up Your Tracking First

Before launching any ads, you must set up proper conversion tracking.

Without it, the platforms cannot learn who to target. This is the section Rex feels most strongly about – it is arguably the single biggest determinant of whether your paid media program succeeds or fails. If your tracking only captures 40% of actual conversions (blocked by ad blockers, iOS privacy changes, or broken pixels), the algorithm is working with a distorted picture of reality.

Modern paid media relies heavily on machine learning. Google’s Smart Bidding, Meta’s Advantage+, and LinkedIn’s ad delivery systems all optimize based on the signals you send back to the platform. When a user clicks your ad and then converts (fills out a form, starts a trial, books a demo), that conversion event tells the platform: “this is the kind of person I want more of.” The platform then uses that signal to find more users who look similar.

The modern tracking stack:

Server-side tracking with Conversions API (CAPI). Browser-side pixels are increasingly blocked by ad blockers, iOS privacy changes, and cookie restrictions. Server-side tracking bypasses all of these limitations by sending conversion data directly from your server to the ad platform’s server. This is no longer optional. It is the foundation of effective paid media in 2026. Google calls it Enhanced Conversions. Meta calls it Conversions API. LinkedIn calls it the Conversions API as well. The implementation is technical, but the impact is dramatic: companies that implement server-side tracking typically see a 15-30% improvement in attributed conversions, which in turn gives the algorithms significantly more data to optimize with.

Advanced Matching. This sends hashed customer data (email addresses, phone numbers) to the ad platforms to improve attribution accuracy. When a user converts, the platform can match that conversion back to the specific ad click more reliably, even across devices and sessions. This is especially important for B2B where buying cycles span weeks or months: the person who clicked your LinkedIn ad on their phone two weeks ago might convert on their desktop today. Without Advanced Matching, that conversion is invisible to the platform.

Consent Mode. With GDPR, CCPA, and evolving privacy regulations, you need tracking that works within consent frameworks. Google’s Consent Mode adjusts data collection based on user consent preferences, allowing the algorithms to still learn from aggregated, anonymized signals even when users decline full tracking. This is especially important if you are running campaigns in Europe or targeting privacy-sensitive enterprise buyers.

Think of it like training a machine learning model. The better the data you provide, the better the outcomes. Without clean tracking, the algorithm is essentially flying blind.

One more critical point: by 2026, platforms like Google (Performance Max, AI Max for Search) and Meta (Advantage+) no longer offer automation as an optional feature. They assume it. The AI handles bidding, audience targeting, creative assembly, and placement. The role of the human operator has shifted from managing campaign settings to ensuring the quality of inputs. Your tracking data, creative assets, product feed, and landing page structure are now the primary levers you control. Everything else is delegated to the machine.

Think of it like training a machine learning model. The better the data you provide, the better the outcomes. Without clean tracking, the algorithm is essentially flying blind.

Step 4: When Something Works, Hit the Gas

Eventually you will find a campaign that produces real pipeline. When that happens, many companies often make another mistake.

They scale too cautiously. The risk of scaling too slowly is often larger than the risk of scaling too fast. CPC has risen for 87% of industries over the past year (WordStream). B2B SaaS CPCs specifically rose ~9% YoY. The window to scale a winning channel at favorable economics is always narrowing.

If a channel produces positive ROI, the best time to invest is now.

Ad inventory rarely gets cheaper over time. Competition increases, prices rise. Winning companies move quickly once a channel proves itself.

Paid media works best when you:

Test aggressively
Identify winners
Scale quickly

Speed matters. Follow this rhythm: test aggressively in months 1-3, identify winning combinations, then shift to scaling mode where 70-80% of budget goes to proven winners and 20-30% stays in testing to find the next breakthrough. Never stop testing entirely.

The 2026 Shift: AI Is Rewriting the Paid Media Playbook

One more dimension worth addressing, because it changes how every founder should think about this channel going forward. The role of AI in paid media has shifted from “nice-to-have optimization” to “the entire operating system.” Understanding this shift is critical for anyone building a paid media program in 2026.

Creative is now the primary lever

When the platform controls bidding and targeting (which it increasingly does, by default), the variable that separates a strong campaign from a weak one is the quality of what you feed the machine.

Platforms are becoming more opaque

Google’s Performance Max and Meta’s Advantage+ give advertisers less control over individual campaign settings. You can not manually set bids for specific keywords in Performance Max. You can not control which placements your Advantage+ ads appear on. The tradeoff is often better aggregate performance at the cost of visibility.

The operator’s role is shifting

Less hands-on-keyboard campaign management. More strategic oversight, creative direction, first-party data management, and measurement architecture. The skill set that matters in 2026 is not bid management. It is understanding what signals to feed the system, how to build a creative testing pipeline, how to own first-party data strategy, and how to interpret why the AI made the decisions it did.

For more on how AI is reshaping software broadly, see our edition on "The “SAASpocalpse.”

Final Thought

Paid media is not magic. But it is one of the most powerful growth engines available to B2B companies when the timing is right and the execution is systematic.

The opportunity is still massive. The best time to build your paid media muscle is before your competitors figure it out. Start now. Start small. But start.

Tag @GTMnow so we can see your takeaways and help amplify them.

Granola raises $125M at a $1.5B valuation, up from $250M less than a year ago. Meeting notes are a commodity at this point, and Granola knows it — the real bet is on becoming the enterprise context layer, with new APIs that let teams pull meeting intelligence into AI workflows. The valuation jump says investors believe whoever owns the context wins.

VC: Investing at Inception in the Age of AI Agents with Ed Sim

For the full thing, listen on Apple, Spotify, YouTube or wherever you get your podcasts by searching “The GTMnow Podcast.”

Examen - raised $4.3M in total funding and launched. Examen is building an autonomous analyst for commercial real estate.

Enterprise Account Executive, US & Canada at Semrush (Dallas, TX)
Customer Success Manager at Esper (Austin, TX)
Customer Success Manager at gaiia (Remote - Canada)
Associate Account Executive (Sports Cards) at Alt (Remote - US/Canada)
AVP, customer success (West) at Writer (Hybrid - San Francisco, CA)
Growth Marketing Manager (PALs/Consumer) at Tavus (San Francisco, CA)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

MicroConf 2026: April 12–14, 2026 (Portland, OR)
GTMfund Dinner: April 14, 2026 (Austin, TX)
HockeyStack’s Top Secret Launch Party: April 16, 2026 (San Francisco, CA)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
CVC Week by Counterpart Ventures: September 29, 2026 (San Francisco, CA)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow community (founder, operator, investor) love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Sophie Buonassisi and the GTMfund team (not AI!).

VC: Investing at Inception in the Age of AI Agents

GTMnow — Wed, 01 Apr 2026 19:08:13 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by: AngelList

They’ve scaled with us across funds and into the future.

If your fund is growing in size or complexity, check them out at https://angellist.com/gtmfund.

Who we sat down with

Ed Sim has been a VC for 30 years. He’s backed companies like Clay, Front, BigID, and Snyk. He writes What’s Hot in Enterprise IT every single Saturday, 489 weeks in a row. And right now, he says this is the most exciting and terrifying moment he’s ever seen in his career.

In this episode, Max and Ed break down what’s actually happening inside startups and boards right now, why the old playbooks are dead, and what separates the companies that will survive this AI shift from the ones quietly getting killed by it.

Discussed in this episode

Why engineering is no longer your bottleneck (and what is)
The 5 P’s Ed uses to evaluate every inception-stage investment
The autonomous enterprise thesis and what it means for how companies are built
Why AI-native leadership is now a survival reqxtuirement, not a nice to have
The full Clay story: $600K to $100M ARR, how they stayed lean, and what actually unlocked growth
The 3 CH’s framework for being a great board partner to founders
Why the best founders today are inside the AI jet stream, not chasing it
What every board meeting sounds like right now

Episode highlights

01:05 – Episode Preview: Ed Sim & Key Takeaways

03:10 – The Jet Stream Analogy: Two Types of Companies

05:43 – How GTM Operators Should Evaluate Companies Like Angel Investors

07:20 – The Collapsing of Moats & AI-Native Business Opportunities

10:00 – Rebuilding Industries vs. Selling Software to Them

15:00 – Why Old GTM Playbooks Are Dead

17:46 – Ed Sim’s Background: From Cutco to 30 Years in VC

21:43 – The Five P’s of Inception Investing

23:04 – How to Evaluate Potential & TAM in a Fast-Changing Market

25:40 – Staying Ahead of the Jet Stream as a Founder

26:32 – The Autonomous Enterprise Thesis

28:44 – Agent of the Week: How Companies Should Adopt AI Agents

29:10 – How Agents Are Changing Engineering Bottlenecks

31:15 – What Incumbents Must Do to Survive the AI Wave

32:53 – Intercom, Snowflake & How Legacy Companies Are Adapting

36:43 – The Clay Story: How They Found Their Footing

38:33 – The Three C’s of Working With Founders (Cheer, Challenge, Chill)

40:07 – Clay’s Growth Trajectory: $600K to $100M+ ARR

41:10 – Clay’s Agency GTM Model & Community Moat

43:50 – Ed’s Fund Model: $500K to $15M Checks at Inception

46:57 – What’s Hot in Enterprise IT & Venture Right Now

48:03 – Closing Remarks

View the Full Transcript

Key takeaways

1. Engineering is no longer your bottleneck.
Your people are. Code is shipping faster than your sales, marketing, and customers can absorb it. The constraint has flipped completely and most companies haven’t noticed yet.

2. Painkillers beat vitamins every time.
The only startups worth backing at inception are solving a hair-on-fire problem someone desperately needs fixed, not a nice-to-have they can live without.

3. The 3 CH’s of being a great board partner.
Know when to Cheer (when founders are getting beaten up), when to Challenge (when they feel invincible), and when to Chill (when they just need breathing room to figure it out). Elliot used all three with Clay to perfection.

4. If your CEO came from sales, you are in trouble.
Surviving this AI shift requires product-driven, agent-native leadership at the top. The companies that adapted, Snowflake, Intercom, Atlassian, all changed leadership first.

5. The best founders are inside the AI jet stream, not chasing it.
The question Ed asks every founder today: are you struggling to keep up, or are you the one constantly shipping and adapting faster than anyone can copy you?

Follow Ed Sim

LinkedIn: https://www.linkedin.com/in/edsim
X (Twitter): https://x.com/edsim
boldstart ventures’ LinkedIn: https://www.linkedin.com/company/boldstart-ventures
boldstart ventures’ website: https://boldstart.vc

Follow Max Altschuler and Paul Irving

Max Altschuler on LinkedIn: https://www.linkedin.com/in/maxaltschuler
Max Altschuler on X (Twitter): https://x.com/HackItMax
Paul Irving on LinkedIn: https://www.linkedin.com/in/paulsirving
Paul Irving on X: https://x.com/PaulGTM
Newsletter:

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

Follow GTMfund

Newsletter: You’re in it :) If this was forwarded to you, you can subscribe here. GTMnow is our main brand for content, but we also have two around our investments:
LinkedIn: https://www.linkedin.com/company/gtmfund
X: https://x.com/gtmfund

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast shares how the best in tech build, scale and invest.
Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

GTMnow is run by GTMfund, an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

The Founder-Led Sales Playbook

GTMnow — Fri, 27 Mar 2026 22:03:38 GMT

Most early-stage founders know their product better than anyone alive. But selling is a different nuanced beast. We recently hosted a private workshop on founder-led sales for 20 founders, and the feedback was incredible.

We partnered with Pete Kazanjy on the event - author of Founding Sales, co-founder of TalentBin (acquired by Monster), Atrium, and Modern Sales Pros.

These are a few of the notes from the session, and that serve as a guide to founder-led sales.

That’s why teams like RingCentral, Outreach, ActiveCampaign, and Fortune 100 companies rely on HockeyStack to eliminate wasted spend, take better decisions, and make space to think. Learn more at hockeystack.com

First, Calibrate What Sales Is

Pete Kazanjy’s definition is simple: Modern sales is consulting where the consultant has a particular predilection to one solution - theirs.

If your job as a salesperson is to consult - to genuinely help a prospect understand their problem, believe in its importance, and evaluate whether your solution is the right fit - then founders are uniquely equipped for it.

Nobody understands the problem you solve better than you do. Nobody has had more conversations with people who have that problem. Nobody can speak with more credibility about the category, the alternatives, the failure modes of the existing approaches.

“Sales is a transfer of enthusiasm.” The founder’s enthusiasm for the problem they’re solving is the single most powerful selling asset in existence. Use it.

The other thing worth internalizing early: companies do not buy things, people do. When you are selling to a business, you are persuading a collection of individuals, each with their own awareness of the problem, their own career incentives, and their own threshold for conviction. The organizational buying decision emerges from the sum of those individual journeys. Understanding this is the foundation of everything that follows.

The Journey Every Buyer Takes

Before you can sell anything, you need a map of where you are taking people. Every buyer, regardless of industry or deal size, moves through a predictable progression of belief. Pete frames it as five stages: Problem Awareness, Problem Prioritization, Solution Preference, Commercial Agreement, and Championship.

Problem Awareness is the starting point — the buyer knows, at some level, that a problem exists. Problem Prioritization is when they decide it is important enough to actually address. Solution Preference is when they conclude that your approach is the right one. Commercial Agreement is the signed contract. Championship is when they become an active internal advocate, willing to fight for the budget, navigate the procurement process, and bring others along.

Your job as a founder-seller is to advance people along this journey. The tools you use — email, phone, a well-run demo, a dinner, a reference call — are the vehicles. But the destination is always the same: a person who started with vague awareness of a problem and ended up a champion of your solution, ready to build org-wide conviction around it.

The most common mistake founders make is pitching solution preference to someone who has not yet prioritized the problem. You can have the most compelling demo in the world, but if the person on the other side of the call does not believe this problem is in their top three priorities right now, nothing is going to happen. Qualifying where someone is on this journey — and meeting them there — is one of the highest-leverage skills in early-stage sales.

The job is not to convince people to buy something they don’t want. The job is to advance people who already have the problem along a journey toward conviction. Start with the right people.

Stage 1: Figuring It Out

The first mode of founder-led sales is the hardest to be honest about, because it requires admitting that you do not yet know things you might assume you already know. You do not know who exactly will buy your product. You do not know what sequence of conversations converts them. You do not know what objections are real versus noise. Stage One is the process of finding out.

The goal of Stage One is not to close as many deals as possible. The goal is to develop a repeatable, transferable understanding of how your product gets sold. Pete calls this the minimum viable sales motion — a small-scale, high-quality proof that you can reliably move a meaningful sample of right-fit prospects from first contact to paying customer at a win rate that starts to feel statistically meaningful. The target is somewhere in the range of 30 to 50 qualified prospects run through your process, with 10 to 20 converting. Once you have that, you have something to build on.

Define your ICP with precision

Everything in Stage 1 starts with Ideal Customer Profile (ICP). Pete breaks ICP down into three nested questions, each more specific than the last.

The first question is what companies: Which organizations have the problem you solve, in an acute enough form that they would pay money to address it? This is where you think about firmographics — industry, company size, growth stage, tech stack, revenue range — but also about structural conditions. What has to be true about a company’s situation for your problem to feel urgent?

The second question is who inside those companies: Which specific individuals feel this problem most acutely and have the standing to do something about it? Title matters, but job function matters more. You are looking for the person whose day-to-day is materially affected by the problem you solve — and who has enough organizational standing to either buy or champion a purchase. That person is your entry point.

The third question is the one most founders skip entirely: when. Is there a specific trigger — a funding event, a new hire, a product launch, a missed quota, a compliance deadline, a tool contract renewal — that makes a company’s need for your solution an urgent priority? Timing is everything in B2B sales. A company that is a perfect ICP fit but is mid-budget-freeze is not a real prospect right now. The companies that are perfect ICP fits and are experiencing a trigger event are your hottest targets. Find those.

Most founders define their ICP too broadly. The single most leveraged thing you can do in Stage 1 is narrow it. The more precisely you can define who has your problem and why, the less time you waste on conversations that were never going to convert.

Build your sales narrative before you build anything else

Before you send a single prospecting email, you need a narrative. This is a clear, coherent story about the problem in the world, why it matters, why it is hard to solve with existing approaches, and why your solution addresses it in a way that nothing else does.

This narrative needs to exist in multiple forms: a spoken version for the first discovery call, a deck for the formal presentation, email sequences for prospecting, a one-pager for the prospect to share internally. The words and framing should be consistent across all of them. When they are not, buyers sense it — and the inconsistency creates doubt about whether you actually understand the problem as deeply as you claim.

The discipline here is writing things down and testing them explicitly. Most founders have a great spoken narrative but have never written it out, which means it lives entirely in their heads and cannot be transferred to a rep, evaluated for weaknesses, or improved systematically. Write down your sales story. Read it out loud. Run it past people who know nothing about your space and see if they understand it. Then run it past people who are in your ICP and see if it resonates. Treat the narrative as a product.

Prospect

Stage 1 prospecting is a deliberate, structured activity. The starting point is a list: a spreadsheet of 75 to 100 target accounts that match your ICP hypothesis, with specific people identified at each — name, title, LinkedIn profile. Building this list is itself a forcing function. If you cannot identify 75 to 100 real companies that fit your ICP, your ICP is either too narrow or too theoretical.

The outreach itself should be human and personal at this stage. This means actual personalized messages that demonstrate you did your homework: you read their LinkedIn, you noticed a relevant trigger event, you understand something specific about their situation. At this stage, the conversion rate on truly personalized outreach dwarfs anything a sequence tool can produce. You are not trying to generate volume yet.

Different buyers respond to different channels. Some will respond to a thoughtful cold email, others will engage on LinkedIn, others need a warm introduction, a few will only meet you at an industry event over a drink. Part of Stage 1 is figuring out which engagement modes work best for your specific ICP, and building that knowledge into the motion you eventually systematize.

The Buying Group Problem

In B2B, there is often no single “decision maker.” In almost every sale, the purchase decision emerges from a group of people who each need to reach a sufficient level of conviction. The executive sponsor, the day-to-day champion, the technical evaluator, the procurement lead, the skeptic who asks the hardest questions in the demo — each of them is on their own journey. Your job is to bring them along together.

Pete calls this dynamic the buying group, and the challenge it creates is one of the most common reasons founder-led deals die in late stages. The founder has done brilliant work building conviction with one person — the champion — but has not gotten adequate access to the other stakeholders who need to be convinced. Then procurement comes in with hard questions nobody has prepared for, or the executive who controls the budget has never heard a real case for the product, or the skeptic surfaces objections that derail three weeks of momentum.

The antidote is deliberate multi-threading. From the first conversation, you should be thinking about who else in the organization needs to be part of this process. Mapping the buying group is the core work of enterprise sales. In simpler sales (lower ACV, smaller buyer organizations), you might have one or two people to convince. In more complex deals, you can have five, seven, ten stakeholders across multiple teams, each needing a different conversation. Understanding which kind of sale you are in — and having a plan for each stakeholder — is necessary.

The deal does not close when one person is convinced. It closes when org-wide conviction reaches a threshold. Your job is to run a coordinated campaign across the entire buying group, not a single sales call with a single champion.

One practical implication: your champion cannot do all this work for you. A champion who loves your product but cannot get internal access or does not have organizational standing is a limited asset. Part of your qualification work in Stage 1 should be assessing the champion’s ability to build internal conviction — their seniority, their political capital, their access to the people who control the budget. A great champion inside a broken buying structure will not close.

Stage 2: Scaling It

The transition from Stage 1 to Stage 2 happens when something specific is true: you have a documented, repeatable sales motion. You can describe exactly what you do from first contact to close, in enough detail that someone else could attempt to follow it. You know which ICP companies to target, which persona to contact, which trigger events to look for, which outreach approach gets responses, how to run the discovery call, what the demo should cover, how you handle the top five objections, what the proof point looks like that moves people from solution preference to commercial agreement.

If you cannot describe all of that, you are not in Stage 2 yet. You are still in Stage 1, and hiring a sales rep right now will cost you time, money, and probably the person’s career momentum. The most common mistake founders make in sales is entering Scale mode before they have actually finished Figure It Out mode.

What needs to be documented before you hire

The handoff from founder to first rep is like a product launch. The product is the sales process itself. Before you bring anyone in to run it, you need a written record of how it works.

That means: a written ICP definition with enough specificity to be operationalizable (not just “mid-market SaaS companies” but the actual firmographic and trigger-event criteria you have validated). A prospecting playbook that specifies the list-building approach, the outreach sequence, and the personalization methodology. A discovery call guide that maps out the questions you ask and what good versus bad answers look like. A demo script that has been tested and refined. An objection guide. A competitive positioning sheet. A typical deal flow with stage definitions that actually reflect how deals progress, not just what a CRM template suggests.

This sounds like a lot. It is. That is the point. If you have not done enough deals to produce all of this from lived experience, you have not done enough deals. The documentation is a forcing function for the self-assessment: do I actually have a repeatable motion, or do I have a collection of things I did that worked for various reasons I do not fully understand? The former can be taught. The latter cannot.

Who to hire first (and who not to)

The first sales hire is one of the highest-stakes decisions in a company’s trajectory. The instinct many founders have is to hire a senior, experienced salesperson — someone who has “been there, done that” and can come in and build something. That instinct is usually wrong.

Here is why: an experienced sales leader hired before the motion is proven will bring their own playbook, which may or may not match your product, your market, or your buyer. They will be frustrated when the infrastructure they expect (clean CRM data, marketing-generated pipeline, refined comp plans) does not exist. They will make assumptions that should still be hypotheses. And when things do not work, it will be genuinely unclear whether the sales motion is broken or the person is the wrong fit.

The right first hire is a high-ceiling, somewhat experienced seller who is analytically rigorous and eager to get into the weeds of a new problem — someone who does not need a polished system handed to them but can follow a documented process and help make it better. Their job is not to invent the sales motion. It is to execute the one you have already proven, generate enough signal to validate that it transfers, and work with you to refine it. Once you have two or three reps running a consistent motion, you have earned the right to bring in sales leadership.

Hire to execute the motion, not to discover it. The founder’s job is to discover and document. The first rep’s job is to validate that it transfers. Leadership’s job is to scale what’s been proven.

For more guidance, check out The First Sales Hire Playbook.

Running the motion efficiently

Stage 2 is about operational discipline. The deals you are closing now need to look like the deals you plan to close at 10x the volume. That means consistent qualification criteria (so you are not wasting time on prospects who were never going to buy). It means a pipeline hygiene standard (so leadership has real visibility into what is actually moving, not what reps are optimistically stage-advancing). It means a deal review cadence that surfaces risk early. And it means a continuous improvement loop — a mechanism for learning from every lost deal and integrating those lessons back into the playbook.

One dimension of Stage 2 that founders consistently underinvest in is the feedback loop between sales and product. In Stage 1, the founder is getting that feedback organically — they hear objections on every call, they see what feature gaps are losing deals, they know what the market wants because they are in the market every day. In Stage 2, as the founder steps back from direct selling, that feedback loop can break. Build mechanisms to keep it alive: regular deal review sessions with reps, close analysis on every lost opportunity, and a formal channel for surfacing product gaps that surface repeatedly in sales conversations.

The Maturity Stages at a Glance

Pete’s framework identifies several distinct maturity stages across the founder-led sales journey. Here is how they map to the activities and questions each stage demands:

Stage 0 — Zero Customers, Zero Motion

You have a product and a hypothesis. The work is: do any of your assumptions about who has this problem and how much they care about it survive contact with actual buyers? Spend 100% of your sales time talking to prospects — not building a deck, not refining the pitch, not setting up a CRM. Get on the phone. Run discovery calls. Listen more than you talk. The goal is validated ICP and problem-pain calibration, not revenue.

Stage 1 — First Customers, No Playbook

You have early customers — maybe five to ten — but you closed them through force of personality, domain credibility, and network effects. You cannot describe exactly how you closed them in a way someone else could replicate. The work is: run 30 to 50 qualified prospects through your process, document every step, track your conversion rates at each stage, and produce a version of the sales motion on paper. By the end of this stage, you should be able to hand a new hire a document that tells them what to do.

Stage 2 — Proven Motion, First Hire

You have a documented motion with statistically meaningful conversion data. A win rate of 15 to 25% on qualified pipeline, a defined sales cycle, and a clear map of the buying group. Now you hire someone to execute it. The early hires are executors, not builders. The critical success metric here is not total ARR — it is whether the rep’s conversion metrics track close to yours. If they do, you have a transferable motion. If they do not, you have more work to do in Stage 1.

Stage 3 — Repeatable Motion, Building Team

Two to four reps are running the motion consistently. Pipeline is flowing from a mix of outbound, inbound, and partner channels. You have enough data to know which deal profiles close and which do not. Now is when you bring in sales leadership — someone who can manage the team, run deal reviews, refine the comp plan, and own the operational infrastructure of the sales org. The founder’s role in sales shifts from doing to enabling.

Final Thought

The most important thing you can take from this edition is the distinction between figuring it out and scaling it.

Nothing happens until someone sells something. That someone, in the beginning, is you. There is no shortcut past it — and there should not be. The founder who has sold is a fundamentally different, more capable leader than the one who never did.

Tag @GTMnow so we can see your takeaways and help amplify them.

Apollo.io acquired Pocus, bringing signal intelligence directly into its GTM platform alongside 230M+ contacts and agentic workflows. The bet: revenue teams shouldn’t need a separate tool to know where to focus. Worth watching how this plays out upmarket, where Apollo has seen 400%+ growth in enterprise accounts over the past year.

GTM 184: How Sophos Scales Customer Success for 600,000 Customers with Teresa Anania

Bonus Episode: GTMfund Hires a New Partner: Jason Demant!

Listen and subscribe wherever you get your podcasts by searching “The GTMnow Podcast.”

Armada - named among Fast Company’s Most Innovative Companies in computing, alongside NVIDIA and Google. The real story is what they’re building: distributed AI infrastructure for the environments most platforms ignore. Worth watching as enterprises and governments push AI into the field, not just the cloud.

Examen - raised $4.3M in total funding and launched. Examen is building an autonomous analyst for commercial real estate. GTMfund is proud to have led the Seed round.

Wyllo - the newly combined NoFraud and Yofi, is reframing fraud prevention as an intent problem rather than a binary approve-or-decline decision. If the thesis lands, risk stops being a cost center and starts informing growth. One to watch for ecommerce teams tired of black-box tools that block as much revenue as they protect.

Enterprise Account Executive at SEMrush (US)
Founding Growth Lead at Casium (Hybrid - Bellevue, WA)
Enterprise Customer Success Manager (Named Accounts) at Gorgias (Hybrid - San Francisco, CA)
Growth Marketing Manager - Demand Generation at CaptivateIQ (Hybrid - Austin, TX | Remote - Raleigh, NC / Nashville, TN / Toronto, Canada)
Enterprise Solutions Engineer (Post-Sales) at Noibu (Remote - Canada)
Senior Manager, Strategic Partner Marketing at Vanta (Remote - US)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

MicroConf 2026: April 12–14, 2026 (Portland, OR)
GTMfund Dinner: April 14, 2026 (Austin, TX)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow community (founder, operator, investor) love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Sophie Buonassisi and the GTMfund team (not AI!).

How Sophos scales customer success for 600,000 customers

GTMnow — Wed, 25 Mar 2026 16:15:23 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Also this week: we welcomed Jason Demant as Partner, Head of Networks at GTMfund. We shared more context in a bonus episode, watch here.

Brought to you by:

If you run go-to-market, you already know the problem: your data lives everywhere. Spreadsheets, CRMs, sales calls, ad platforms… yet you’re still guessing what to do next.

HockeyStack is the AI platform for modern GTM teams. It unifies all your sales and marketing data into a single system of action. Built-in AI agents help teams prospect the right accounts, improve conversions, close and expand deals, and scale what works. That’s why teams like RingCentral, Outreach, ActiveCampaign, and Fortune 100 companies rely on HockeyStack to eliminate wasted spend, take better decisions, and make space to think.

Learn more at hockeystack.com

A clear shift is underway in enterprise software. More companies are rethinking how they segment customers and structure servicing models — especially in post-sale motions like renewals, winbacks, and add-ons.

For too long, too much lower-touch revenue has been routed through direct sales or partner motions where the cost-to-serve doesn’t match the value delivered. Increasingly, companies including Sophos, SUSE, and PTC are using Cleverbridge to move more routine transactions into digital, buyer-friendly purchasing paths while reserving high-touch resources where they add the most value.

If you’re rethinking customer segmentation, servicing models, or how digital and AI should apply across your customer base, Cleverbridge is worth a look. Learn more at grow.cleverbridge.com/lp/digital-commerce.

Who we sat down with

This episode was recorded prior to Teresa Anania’s move to Chief Customer Officer at Verint, another Thoma Bravo portfolio company. At the time of recording, she was SVP of Customer Experience at Sophos.

Teresa Anania (CCO at Verint, formerly CCO at Sophos) joins GTMnow to share how she’s built customer success into a true revenue engine at a company serving 600,000 customers across over $1 billion in annual revenue, and why the old reactive, relationship-based CS model is no longer cutting it.

At Sophos, the threat landscape is compounding fast. AI is accelerating the speed and sophistication of attacks, which means response times, customer journeys, and success motions all have to evolve in lockstep in order to keep up. Teresa has spent her career at companies like Zendesk, Autodesk, and ON24 building the operational frameworks that make that possible at scale.

Discussed in this episode

Why AI-accelerated threats have made “always on” customer operations a baseline requirement
How Sophos structures customer success to respond before a customer even knows there’s an issue
The attribution model Teresa uses to tie CS touchpoints directly to retention and expansion
Why waiting for a support ticket is too late
The crawl, walk, run approach for early-stage companies that want to build CS rigor without perfect data
How dynamic segmentation (not just company size) determines coverage models and CSM assignment
What “humble confidence” looks like as a hiring bar and why Teresa uses it across her org
Why retention has to be an all-company play
The role of inner and outer feedback loops in turning NPS data into cross-functional action
How to start the digital journey transformation without disrupting the core business

Episode highlights

00:00 – Sophos at a glance: 600K customers, $1B+ in revenue, and why cybersecurity keeps growing when everything else slows down

01:51 – AI vs. AI: how threat actors are using the same tools Sophos is building against

03:02 – The 2026 Active Adversary Report: why attackers are logging in, not breaking in

04:20 – Attackers move in 3-4 hours and how Sophos structures CS to respond before the customer even knows there’s a problem

06:05 – Connecting CS activity directly to retention and expansion: the attribution model

08:45 – Advice for early-stage companies that want this kind of rigor but don’t have perfect data yet

10:06 – Automation as a scale lever: crawl, walk, run and why you should start at the end of the renewal cycle

13:43 – The future of go-to-market: self-serve from first touch through win-back, powered by AI

18:20 – “The customer should never feel your org chart”: building a digital journey that meets people where they are

20:39 – Going from legacy manual to digital without blowing up the business

22:00 – Dynamic segmentation: why hard lines on ACV are the wrong way to assign CS coverage

26:02 – The two-by-two that actually matters: risk, spend, and growth potential

32:08 – The humble confidence hire: why Teresa looks for this specific combination across her entire org

34:15 – The 5-to-1 scorecard and what Teresa has learned about earning customer trust over time

36:14 – Inner and outer feedback loops: how Sophos turns NPS data into cross-functional action

38:11 – Why retention has to be an all-company meeting, not a CS slide

View the Full Transcript

Key takeaways

1. If you’re waiting for a ticket, you’ve already failed your customer.
By the time a customer inbounds a problem, you’ve already missed a window. The best CS orgs are predictive. They identify where customers get stuck, build curated digital journeys around those moments, and intervene before the customer even knows they need help. That’s what “always on” actually means.

2. CS only drives revenue when it’s outcome-based, not relationship-based.
The old model (check-ins, QBRs, account health scores) isn’t cutting it anymore. Teresa builds attribution models that correlate specific engagement touchpoints to retention and expansion. You can’t improve what you can’t measure, and most teams aren’t measuring the right things.

3. Dynamic segmentation beats static ACV tiers every time.
Hard lines on company size or contract value lead to misallocated coverage and missed growth. The better model? A two-by-two of current spend, risk, and expansion potential that flexes every few months as customer health changes. Your best CSMs should be matched to the accounts with the most upside, not just the biggest logos.

4. Don’t just treat retention as a CS swim lane.
Teresa runs a monthly cross-functional meeting with product, marketing, sales, and support to action NPS insights together. The inner loop closes individual feedback fast. The outer loop informs strategy. Most companies say retention is everyone’s job. Few actually set it up that way.

5. Start at the end of the renewal cycle, not the beginning.
When building a digital CS motion from scratch, Teresa starts with win-back and renewal, not onboarding. That’s where you can show near-term impact, identify churn culprits, and build the playbooks that eventually get you to a proactive, full-lifecycle motion. Crawl, walk, run.

Follow Teresa Anania

LinkedIn: https://www.linkedin.com/in/teresa-anania
Sophos’ LinkedIn: https://www.linkedin.com/company/sophos
Sophos’ Website: https://www.sophos.com
Verint’s LinkedIn: https://www.linkedin.com/company/verint
Verint’s Website: https://www.verint.com

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast tells the stories of how the top 1% of founders, operators, and investors build, scale and invest.

Each week, we uncover the pivotal moments, bold decisions, and go-to-market strategies that turn ideas into breakout companies.

GTMnow is the media brand of GTMfund - we are an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

Claude for your personal productivity

GTMnow — Fri, 20 Mar 2026 12:05:45 GMT

The first Claude Code edition we published gave you 30+ use cases across every GTM function. Beyond the core functions, a few fun wild ones were included. Since that edition, we’ve been getting a flood of questions about broad-reaching personal productivity use cases, in addition to GTM ones. So that’s what what this edition is all about!

This edition breaks down 8 specific builds you can copy. Each one includes the exact setup, the workflow, and the prompt to get started. Plus, an example of someone who has built it. Pick one and try it tomorrow. Then try another.

Before You Build Anything: The 10-Minute Foundation

Every single person who shared their build emphasized the same starting point: Context. It’s also the number one cited thing from AI experts on what people could do to improve their output.

Claude is only as good as the context you give it. Without context, you get generic output.

The foundation is a single file called CLAUDE.md. Create a folder (~/my-productivity or whatever you prefer), and inside it, create this file with four sections:

Who you are. Your role, company, team size, what you are responsible for. Think of this like a briefing: “VP of Sales, Series B SaaS, 12-person team, $8M ARR, selling to mid-market” gives Claude infinitely more context than “sales leader.”

How you work. Your communication style, preferences, and quirks: “Short sentences. No exclamation marks in professional emails. I end every meeting by restating action items. I prefer Slack for internal and email for external. My reports send updates on Monday mornings.” Think of this like the nuances of how you operate.

What matters right now. Your top 3 to 5 priorities this quarter. This gives Claude a filter for what is urgent versus what is noise: “Close the Acme enterprise deal by March 30. Hire a senior AE by end of Q1. Reduce sales cycle from 45 to 35 days. Launch the partner program pilot.” Update this weekly. It changes your output quality immediately.

Key people. The 10 to 15 people you interact with most. Their names, roles, and how they prefer to communicate: “Sarah (CEO): prefers short Slack DMs, never email for internal stuff. Marcus (CRO): loves data tables, hates narrative paragraphs. Jordan (top AE): responsive on Slack after 3pm, prefers voice notes.” This lets Claude adapt tone per recipient when drafting communications.

This file takes 10 to 20 minutes to write. It is the single highest-leverage thing you can do. The more context Claude has, the better every build performs.

The real skill needed here is learning to articulate your own workflows precisely enough that a machine can execute them. The CLAUDE.md forces you to be explicit about how you actually work. That exercise, even before Claude does anything, is valuable.

Build 1: The Morning Briefing

What it does: Pulls your calendar, inbox, meeting notes, and open commitments into a single, prioritized daily briefing you can scan in 2 minutes.

The setup: Connect Claude to Google Calendar, Gmail, and Granola (or your meeting notes tool of choice). In Claude Code, this happens through MCP server connections. In Cowork, these are built-in connectors you enable in settings. If you use Otter, Fireflies, or another transcription tool, export your transcripts as text files into a dedicated folder.

The prompt:

“Run my morning briefing. Pull my calendar for the next 24 hours. For each meeting, include who is attending, what we discussed last time we met (check recent transcripts), and any prep I should do. Check my Gmail for emails from the last 12 hours that need a response, prioritizing emails from [list your VIPs: CEO, key clients, board members]. Check my recent meeting transcripts for commitments I made that I have not yet acted on, specifically phrases like ‘I will send,’ ‘I will follow up,’ ‘let me check on that,’ or ‘I will get back to you.’ Give me: (1) Today’s schedule with context for each meeting, (2) the 3 most urgent emails and why they are urgent, (3) open commitments from recent meetings with their original deadline, (4) a suggested focus plan that protects at least one 90-minute deep work block.”

Why it works: Most people start their day by opening their inbox and reacting to whatever landed on top overnight. The morning briefing flips that. You start with a synthesized view of what actually matters. It surfaces commitments you forgot you made. It flags the email from your CEO that is buried under 40 newsletters. It tells you that your 2pm meeting is with someone you promised a follow-up to last Tuesday.

People are running this as an automated Claude Code start hook. When you open Claude Code in the morning, the briefing runs automatically. No prompt needed. The briefing is waiting for you. For most people, running it manually with a single prompt is enough, but if you want to take it a step further you could give that a try.

Examples of people who do this: Jim Prosser (open-sourced on GitHub, March 2026), Crystal Widjaja (published on Substack), Ankita Tripathi (published on Medium, January 2026)

Build 2: The Email Triage and Draft Agent

What it does: Reads your inbox, classifies every email into categories (respond, FYI, action needed, skip), and drafts replies that sound like you wrote them. The non-negotiable rule we recommend: it saves drafts only. It never sends.

The setup: Connect Claude to Gmail via MCP server or Cowork connector. Make sure your CLAUDE.md has a section on communication style with real examples of emails you have sent.

The prompt:

“Read the 25 most recent emails in my inbox. For each one, classify it into one of four categories: (1) NEEDS RESPONSE: draft a reply in my voice, matching the tone and formality of the thread. Reference context from recent meetings or past emails if relevant. Save as a draft in the sender’s thread. (2) FYI ONLY: summarize in one sentence. (3) ACTION REQUIRED: extract the specific action I need to take, who is asking, and the deadline. (4) SKIP: newsletters I do not read, automated notifications, cold outreach, spam. For any email involving pricing, contracts, legal matters, hiring decisions, board communications, or anything relationship-sensitive, draft a response but flag it with [HUMAN REVIEW] and a note explaining why this needs my personal attention before sending.”

Why it works: The average person receives 120+ emails per day. Most people spend hours on email, and most of that isn’t writing - it’s deciding what to do with each message, recalling the context of the thread, and matching the appropriate tone. Claude handles all of that. You just review the drafts and hit send on the ones that are ready.

Examples of people who do this: Harper Reed (published on his blog, December 2025), Jim Prosser (open-sourced)

Build 3: The Meeting Prep Machine

What it does: Before any meeting, it pulls your history with that person or company: past meeting transcripts, recent email threads, their company’s latest news, and any open commitments from either side. Delivers a one-page briefing so you walk in fully prepared.

The setup: This build requires Granola (or another meeting notes tool) connected to Claude.

The prompt:

“I have a meeting with [Name] from [Company] at [time]. Prepare a meeting brief. Pull: (1) the last 3 meeting transcripts where this person was an attendee, (2) any email threads between us from the last 30 days, (3) search the web for their company’s latest news, press releases, and blog posts from the last 60 days, (4) check if there are any open commitments from either side in previous meeting notes. Give me a one-page brief with: a summary of our relationship and recent interactions, the key topics from our last conversation, any open commitments, their company’s recent strategic moves, and 3 strong opening questions I can ask based on what is happening at their company right now.”

Why it works: Walking into a meeting without context is walking in with a disadvantage. Prep can take a while: 15 minutes of digging through old emails, searching Slack, trying to remember what was discussed. This compresses that to a 30-second read. The opening questions are particularly powerful because they signal that you pay attention and care about their business.

Examples of people who do this: Crystal Widjaja (Granola + Obsidian integration), Cortney Hickey (Zapier’s EA), Reid Robinson (Zapier’s Lead PM for AI).

Crystal’s approach: She built a custom Granola sync script (Python, open for anyone to modify) that auto-exports meeting transcripts into her Obsidian vault. Claude Code reads the vault, the calendar, and Gmail together. The context is unified. Every meeting prep draws from the full history, not just what you can remember.
Reid’s approach: He uses Claude with Zapier’s MCP integration for strategic intelligence gathering before events. Before attending industry conferences, he asks Claude to research every attendee using Zapier’s CRM and web search capabilities. He walks in with instant context on every person he is meeting.

Build 4: The Post-Meeting Follow-Up Engine

What it does: After a meeting ends, it reads the transcript, extracts every commitment, action item, and decision, and drafts a follow-up email confirming everything.

The prompt:

“Read the Granola transcript from my meeting that just ended with [Name/Company]. Extract: (1) every commitment I made, with the exact quote where I said it and the deadline if one was mentioned, (2) every commitment they made, (3) decisions that were finalized during the call, (4) topics that were raised but deferred or need follow-up, (5) any next meeting or checkpoint that was discussed. Draft a follow-up email that: opens with a thank-you, summarizes the 2 to 3 key outcomes, lists all action items with owners and deadlines, and proposes a next step. Match the tone to the relationship (formal for new prospects, casual for existing clients). Save as a draft.”

Why it works: The gap between “great meeting” and “nothing happened” is almost always a missing follow-up. Research from Gong shows that deals with a follow-up email sent within 24 hours close at significantly higher rates than those without. This build closes that gap in 60 seconds after the meeting ends. And critically: when tomorrow’s morning briefing runs, it will check whether you have acted on those commitments. The builds feed each other.

Examples of people who do this: Multiple operators across the GTMfund community.

Build 5: The Weekly Review

What it does: Steps back from daily execution and gives you a strategic view of your week. What you accomplished, what fell through the cracks, how your time broke down. And one specific coaching recommendation.

The prompt:

“Run my weekly review. Analyze the past 7 days: (1) list every meeting I had with a one-line summary of the outcome or key decision from each, (2) list all commitments I made this week from meeting transcripts and emails, and flag their current status: completed, in progress, or overdue, (3) calculate how my calendar time broke down by category: external meetings vs. internal meetings vs. focus time vs. admin vs. unscheduled time, (4) analyze my email volume: how many emails did I send vs. receive, and what percentage were proactive (I initiated) vs. reactive (I responded). Give me: a one-paragraph executive summary of what I accomplished this week, the 3 biggest open items carrying into next week, any patterns I should be aware of (recurring time wasters, growing commitment backlog, declining focus time), and one specific recommendation for how I should structure next week differently.”

Why it works: Most people’s weeks happen to them. The weekly review is the antidote. It shows you patterns invisible in real-time. Are you spending 70% of your time in meetings and wondering why strategic work never gets done? Are your commitments piling up faster than you clear them? Are you sending 80% reactive emails and only 20% proactive ones?

After a few weeks of running these reviews, the coaching recommendations become genuinely valuable. “This is the third week where your Wednesday afternoon focus block was interrupted by unscheduled calls. Consider blocking that time as a recurring calendar hold.”

Examples of people who do this: Multiple operators across the GTMfund community.

Build 6: The Personal CRM

What it does: Builds and maintains a lightweight relationship tracker from your email and meeting history. Surfaces who you are losing touch with and who you should re-engage.

The prompt:

“Scan my Gmail sent folder and Granola meeting transcripts from the last 90 days. For every external contact I have interacted with more than twice, create an entry with: Name, Company, Role (if identifiable from email signature or meeting context), last contact date, key topics discussed across all interactions, relationship strength (high/medium/low based on frequency and depth of interaction), and any open action items from our last exchange. Flag anyone I have not contacted in 30+ days who has a ‘high’ relationship strength. Also flag anyone who I have had 3+ meetings with but have not emailed in 2+ weeks. Export as a CSV file and also save as a formatted markdown file.”

Why it works: The average person’s network decays at a rate of roughly 5% per quarter when not actively maintained. This build creates the relationship layer that most people are missing. Run it monthly. The “fading relationships” flag alone has been called the most useful output by operators who use it, and a pro tip is connecting other places of communication like DMs.

Examples of people who do this: Multiple Pre-Seed and Seed founders in the GTMfund portfolio.

Build 7: The Voice Notes to Action Items Pipeline

What it does: Takes raw voice notes recorded during walks, commutes, or between meetings and turns them into organized themes, drafted content, and specific action items.

The prompt:

“Read all voice note transcripts in /voice-notes from the past week. These are raw, unstructured recordings of my thinking. For each one: (1) extract any action items or commitments I mentioned (things I said I need to do, people I said I need to contact, ideas I said I should explore), (2) identify the main ideas, themes, or insights. Then: (a) group the themes into coherent clusters, (b) for the most compelling theme, draft a 500-word article or memo that captures the insight clearly, (c) create 2 LinkedIn posts based on the best ideas, matching my writing style from CLAUDE.md, (d) compile all action items into a single prioritized list with suggested deadlines. Save each output as a separate file in /drafts.”

Why it works: Most people’s best thinking happens away from a screen. Morning walks, commutes, the shower, between meetings. This helps to ensure your walking ideas turn into value, however you end up using them.

Examples of people who do this: Helen Lee Kupp (published in Lenny’s newsletter).

Build 8: The File System Cleanup and Organization

What it does: The simplest build on this list and the best place to start if you are new to Claude Code. Claude organizes your messy folders: Downloads, Desktop, Documents.

The prompt:

“Analyze my Downloads folder. For every file, identify: file type, apparent project or topic (based on filename and content if readable), file size, and last modified date. Propose a folder structure that organizes everything logically by: (a) file type (Documents, Spreadsheets, Images, Videos, Code, Archives), (b) project or topic where identifiable, (c) a subfolder called ‘Large Files Over 100MB’ for anything big, and (d) a subfolder called ‘Older Than 6 Months’ for anything that has not been accessed recently. Show me the proposed plan with file counts per category and estimated storage savings before moving anything. After I approve, execute the reorganization, rename any files that have unclear names (like ‘IMG_4782.jpg’ or ‘Untitled-3.pdf’) to something descriptive based on their content, and create a summary log of every file moved with its old and new location.”

Why it works: Super easy and rewarding to see messy files organized. This is the build that gives people the “aha moment” when they’re getting started. It is also legitimately useful: most people have 200+ files in their Downloads folder accumulated over months or years (guilty!). You’ll also likely free up lots of storage from duplicate files, obsolete downloads, and ancient screenshots they forgot existed. It is the best on-ramp to Claude Code because you see results immediately, the risk is near-zero (Claude shows you the plan before executing), and the satisfaction of a clean file system is immediate.

Examples of people who do this: Dan Shipper (CEO of Every), multiple operators in the first GTMnow Claude Code edition.

The Pattern Across All 8 Builds

If you look at what these operators built, a clear pattern emerges across every system:

1. They all start with context, not prompts. The CLAUDE.md file, the Granola transcripts, the email history, the sent folder patterns. Every operator emphasized that output quality is directly proportional to context quality. The prompt is the last thing you write. The context is the first.

2. Keep humans in the loop at critical moments. Nobody should be letting Claude send emails or make decisions unsupervised. The system drafts, the human decides. Claude triages, the human prioritizes. Claude flags commitments, the human decides which ones to honor. As I like to say: “don’t outsource your judgment.” It’s about eliminating the 80% of work that is information retrieval, context assembly, and formatting. The 20% that is strategy and relationships stays human.

3. They all compound over time. The morning briefing catches commitments extracted from yesterday’s post-meeting follow-up. The email draft agent uses meeting context from the prep machine. The weekly review spots patterns across all the other builds. Each system feeds the others, and the more you use them, the richer the context becomes and the smarter the outputs get. This is the fundamental difference between a single prompt and a system.

The GTMfund team has a Claude and AI working session on Friday next week, so undoubtably there will be more Claude editions coming soon 🤖

Tag @GTMnow so we can see your takeaways and help amplify them.

The State of GTM Engineering report is out. Worth a read if you’re rethinking how your revenue engine is built.

Bobby Morrison, former CRO of Shopify, published 10 Practical Lessons From Building an AI-Native Company and the core argument is: AI fails when treated as a technology initiative instead of an operating model. Democratized experimentation beats centralized governance every time. The section on managers is worth the read alone.

VC Episode with Brett Queener What wins when anyone can build anything?

GTM 183: Inside ServiceNow’s $10B GTM Engine with Paul Fipps

Listen and subscribe wherever you get your podcasts by searching “The GTMnow Podcast.”

Examen - raised $4.3M in total funding and launched. Examen is building an autonomous analyst for commercial real estate. GTMfund is proud to have led the Seed round.

Gumloop - raised a $50M Series B from Benchmark. They are bringing no-code AI agent building to every employee, not just engineers.

Senior Customer Solutions Manager at Esper (Hybrid - Austin, TX)
Senior Manager Demand Generation (ABM & Revenue Marketing) at Vividly (Remote - US)
Head of Sales, Enterprise & Strategic at Vanta (Remote - US)
Account Executive, SMB at Gorgias (Hybrid - Paris, France)
Strategic account executive at Writer
1. Central (Hybrid - Chicago, IL; Austin, TX; Dallas, TX; Indianapolis, IN; Minneapolis, MN)
2. East (Hybrid - New York City, NY; Boston, MA; Hartford, CT; Jersey City, NJ; New Haven, CT; Newark, NJ; Philadelphia, PA; Stamford, CT)
3. West (Hybrid - San Francisco, CA; Denver, CO; Los Angeles, CA; Seattle, WA)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

Spryng (for marketers): March 24–25, 2026 (Austin, TX)
MicroConf 2026: April 12–14, 2026 (Portland, OR)
GTMfund Dinner: April 14, 2026 (Austin, TX)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow community (founder, operator, investor) love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Sophie Buonassisi and the GTMfund team (not AI!).

GTM: From 900 killed AI pilots to $335M in gains

GTMnow — Wed, 18 Mar 2026 16:22:58 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by: HockeyStack

If you run go-to-market, you already know the problem: your data lives everywhere. Spreadsheets, CRMs, sales calls, ad platforms… yet you’re still guessing what to do next.

Learn more at hockeystack.com

Who we sat down with

Paul Fipps (President of Global Customer Operations at ServiceNow) joins GTMnow to break down how ServiceNow built the customer engine behind $10B+ in revenue and 20%+ growth for five consecutive years.

From CIO at Under Armour overseeing a 300 million-member connected fitness ecosystem, to now leading global sales, customer success, field marketing, and partners at one of the most disciplined GTM organizations in enterprise software, Paul has seen what it takes to scale from both sides of the table.

Discussed in this episode

Why complacency is a bigger threat than competition at scale
How to detect churn long before it shows up in a report
What a CIO cancelling 900 AI pilots tells you about where enterprise AI is actually headed
How ServiceNow unified sales, customer success, field marketing, and partners into one GTM motion so customers never feel the org chart
Why ServiceNow monitors customer health daily — and what signals their teams actually track
How community became a core GTM advantage, not just a marketing channel
How ServiceNow’s AI Control Tower governs agents across the enterprise stack
Inside “Now on Now”: how ServiceNow generated $335M in annualized AI productivity gains using its own platform
How integrating Claude into the GTM workflow cut account planning from days to minutes
What DTC product thinking from Under Armour unlocked in enterprise GTM
How ServiceNow shifted from 6-month product releases to monthly innovation cycles
Paul’s advice for building a world-class GTM organization: put the best people in the right seats

Episode highlights

00:00 – How ServiceNow built one of the most disciplined GTM engines in enterprise software

01:22 – 80B workflows, $10B+ revenue: what gets harder and easier at scale

02:05 – Why complacency is the real threat at scale

05:49 – Why ServiceNow unified sales, customer success, field marketing, and partners under one motion

06:49 – The post-sale handoff problem: signing on Friday, new team showing up Monday

08:22 – How to spot churn before it shows up in a report

09:55 – How often ServiceNow teams check customer health

12:25 – If you took away the dashboards, how would you know a customer is truly winning?

15:35 – Why Paul blocks calendar time every week for direct customer conversations (and responds within 24 hours)

17:53 – From Under Armour to ServiceNow: what DTC product thinking unlocks in enterprise B2B

21:06 – The personalization gap in B2B enterprise software and how ServiceNow is closing it

25:18 – The CIO with 900 AI pilots who cancelled every single one

26:03 – How embedding agentic AI inside existing workflows drives measurable ROI

32:39 – “Now on Now”: $335M in productivity gains running on their own platform

34:18 – Integrating Claude into the GTM motion for all 10,000 go-to-market team members

36:44 – The AI control tower: governing every agent across the entire enterprise

39:15 – Paul’s one piece of advice for every GTM leader: get the best people in the right seats

40:21 – The book that shaped Paul’s career: Execution by Bossidy & Charan

View the Full Transcript

Key takeaways

1. At scale, complacency is a bigger threat than competition.
When you’re growing 20%+ year over year, the danger isn’t who’s coming to eat your lunch. The internal drift that happens when you’re adding thousands of people and hundreds of new processes to your org is a much greater existential threat. Paul’s job is to make sure every single person stays oriented around one thing: making the customer the hero of their own story.

2. Customers should never feel your org chart.
ServiceNow unified global sales, customer success, field marketing, and partners under one motion, deliberately. When you do all the hard work of a complex enterprise sale and then an entirely new team shows up Monday morning asking, “so what are we doing?” you’ve already lost. The pre-sale and post-sale have to have one continuous through line.

3. Churn shows up in relationships long before it shows up in reports.
They monitor adoption daily. But the signal Paul watches most closely is behavioral: when a customer stops showing up to QBRs, when they disengage from meetings, etc - you know something is already wrong. The best time to act on churn is before any data shows it’s an issue.

4. The AI governance crisis is real and it’s happening right now.
Paul spoke with the CIO of one of the largest healthcare companies in the world. The CIO had 900 AI pilots running inside his organization, yet he cancelled every single one. The security and governance risk outweighed the innovation benefit. For every company racing to deploy AI: can you govern pilots as fast as you can deploy them?

5. Consumer product thinking is still an unlock in enterprise B2B.
Most enterprise software companies hire from within enterprise software. Paul brought a different lens from Under Armour where data from 300 million users surfaced unarticulated needs that led to product breakthroughs. In B2B, that same instinct to treat every buyer as an individual with unique needs, not just a segment or a persona, is still a meaningful competitive advantage most companies haven’t fully captured.

Follow Paul Fipps

LinkedIn: https://www.linkedin.com/in/paulfipps
ServiceNow LinkedIn: https://www.linkedin.com/company/servicenow
ServiceNow website: https://www.servicenow.com

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast tells the stories of how the top 1% of founders, operators, and investors build, scale and invest.

Each week, we uncover the pivotal moments, bold decisions, and go-to-market strategies that turn ideas into breakout companies.

GTMnow is the media brand of GTMfund - we are an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

VC: What wins when anyone can build anything?

GTMnow — Tue, 17 Mar 2026 17:14:23 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by: AngelList

They’ve scaled with us across funds and into the future.

If your fund is growing in size or complexity, check them out at https://angellist.com/gtmfund.

Who we sat down with

Brett Queener (Partner at Bonfire Ventures) joins GTMnow to share what three decades across Siebel, early Salesforce, co-founding, and seed-stage investing has taught him about what actually wins now that software is cheaper and faster than he ever imagined.

Brett was one of the earliest GTM hires at Salesforce when it had seven employees. He helped build the go-to-market playbook that defined a generation of SaaS: enterprise segmentation, sales motion design, product marketing, the whole works. He then co-founded SmartRecruiters, angel invested in companies like Outreach and Pando, and eventually joined Bonfire Ventures as a Partner to do early-stage investing the way he thinks it should be done: hands-on, operator-led, and built around founders who are ruthless about execution.

Discussed in this episode

Why “the king of explaining” is no longer a competitive advantage in go-to-market
How AI collapses the traditional sales motion from SDR qualification through CSM QBR
Why your right to win becomes table stakes every 30 days and what to do about it
What Brett looks for in founders now, and why he calls the best ones “psychos”
Why vertical software is more defensible than ever in the AI era
Why face-to-face still matters when buyers are betting their careers on a product that didn’t exist six months ago
How the Salesforce and HubSpot ecosystems could become their own boat anchors
Why events are creating 75%+ of pipeline for Brett’s early-stage portfolio companies
What “deploy before you close” looks like in practice
Why the functional GTM leader of today has to be a builder, not just a smooth talker
What changes in VC when building software no longer requires capital or headcount to get started
Why seven out of ten bets should die and what that means for how seed-stage investors should think about portfolio construction

Episode highlights

00:00 – Brett’s career: Siebel, Salesforce employee #7, co-founder, angel investor, seed-stage VC

06:58 – The big shift: from passive CRUD apps to agentic software that does the work for you

08:07 – “Failing upwards” and what the fastest-growing companies taught Brett about investing in people

10:58 – Why Brett moved into VC: 25 years of operators whose collective worth hit $25B

16:50 – What Brett looks for in founders now and how it’s changed over 6-7 years

18:05 – Why anybody can build anything: Brett builds a fully functional travel app in 15 minutes

19:20 – The last remaining moat in software

26:20 – The real threat to Salesforce and HubSpot and why their ecosystem might be the boat anchor

34:20 – How the entire GTM motion changes when the product does the job instead of just enabling it

38:30 – Why communicating the right problem to the right ICP hasn’t changed in 30 years

39:20 – Deploy first, close later

40:20 – Why face-to-face matters more now, not less

41:00 – Why events are driving 75%+ of pipeline at Brett’s early-stage portfolio companies

42:10 – Pricing and packaging: the most important GTM lever nobody talks about enough

46:35 – The death of the smooth-talking GTM leader

47:25 – What happens to VC when software is no longer scarce to build

52:00 – Why vertical software wins in the AI era: context, workflow, and the non-tech buyer

55:20 – The rep who drove 65 miles to drop off donuts and closed an $80K deal

View the Full Transcript

Key takeaways

1. The GTM playbook most of us learned is already gone.
The entire motion from SDR to CSM was designed around one problem: the product didn’t do the job, so humans had to explain it. When the product actually does the job, that whole layer of explanation collapses. You’re showing product earlier, deploying before you close, and tracking outcomes from day one. The people who built careers on translating software into value have to find a new edge.

2. Your right to win becomes table stakes every 30 days.
In SaaS, you could build a wedge, raise a Seed round, hit $1M ARR, and spend 18 months turning that into a go-to-market machine. That’s a thing of the past. The rate of change in what your product does (and what your competitors can ship) means the messaging, positioning, and value prop have to move in lock step.

3. The founder profile that wins now looks inherently different.
Brett looks for founders who are building, not managing people who build. You need to be actually in the tools, shipping, and learning what the product can do firsthand. If you’re not building, how can you expect to have a clear enough point of view on what you’re selling or how to sell it?

4. Face-to-face matters more in the AI era.
Buyers right now are more nervous than ever. They’re betting their careers on products that are evolving fast and vendors who might look completely different in 6 months. They’re buying whether they trust you to build what they’ll need tomorrow. That trust is still built in person. It’s no surprise that events are creating 75%+ of the pipeline for Brett’s early-stage portfolio companies.

5. Vertical software is the most defensible bet in AI right now.
In vertical markets, customers hand you everything: their workflow, their edge cases, their entire context. That data is the moat. And to make matters even better, the non-technical buyer is now one of the most attractive customers in the market because AI finally makes their job meaningfully easier without requiring them to learn anything.

Follow Brett Queener

LinkedIn: https://www.linkedin.com/in/brettqueener
Bonfire Ventures’ LinkedIn: https://www.linkedin.com/company/bonfireventures/
Bonfire Ventures’ website: https://www.bonfirevc.com

Follow Max Altschuler and Paul Irving

Max Altschuler on LinkedIn: https://www.linkedin.com/in/maxaltschuler
Max Altschuler on X (Twitter): https://x.com/HackItMax
Paul Irving on LinkedIn: https://www.linkedin.com/in/paulsirving
Paul Irving on X: https://x.com/PaulGTM
Newsletter:

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

Follow GTMfund

Newsletter: You’re in it :) If this was forwarded to you, you can subscribe here. GTMnow is our main brand for content, but we also have two around our investments:
LinkedIn: https://www.linkedin.com/company/gtmfund
X: https://x.com/gtmfund

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast tells the stories of how the top 1% of founders, operators, and investors build, scale and invest.

Each week, we uncover the pivotal moments, bold decisions, and go-to-market strategies that turn ideas into breakout companies.

GTMnow is the media brand of GTMfund - we are an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

The Growth OS Map: Winning in the Age of Channel Collapse

GTMnow — Fri, 13 Mar 2026 21:05:30 GMT

This edition felt like a particularly important one, given the speed at which channels are evolving.

TL;DR

The problem: CAC payback has hit 57 months. Traditional channels (SEO, Outbound, Ads) are collapsing under AI noise.
The shift: You can no longer rent attention. You must build an Operating System of defensible “Growth Loops” that compound over time.
The solution: This guide maps the 8 specific loops (Viral, Content, Community, etc.) and the 5-layer architecture needed to replace fragile funnels with a resilient GTM engine.

This guide was created in collaboration with Stefan Bader, co-founder of Cello, a platform helping SaaS companies build and scale referral-driven growth loops.

The End of Renting Distribution

We all know the playbook. For fifteen years, it was simple: raise a Series B, hire an army of SDRs, light money on fire with Google Ads, and force-feed the top of the funnel.

That playbook is dead.

Building a product used to be the hard part. Now, with AI tools like Anything, Lovable and Cursor, shipping a functional SaaS MVP takes hours, not months. Product reproducibility has become trivial. The new bottleneck is distribution.

The numbers don’t lie. According to Jamin Ball’s Clouded Judgement, the median CAC payback period for public SaaS companies has ballooned to 57 months. That’s nearly five years just to break even on a customer.

Look at Reddit. Their stock dropped over 16% in a single day when ChatGPT reduced Reddit citations from 14% to just 2% of responses. That’s the risk of building your house on land you don’t own.

We call this the AI Product-Channel Fit Collapse. Your product might be incredible. But it’s your distribution that determines your rate of success now.

The AI Product-Channel Fit Collapse

Traditional channels are breaking. Here is the reality across the board:

1. SEO and Content: Traffic Without Clicks

Similarweb data shows zero-click searches jumped to 69% in 2025. Meanwhile, AI has flooded the internet. When everyone can publish infinite content, it gets much harder to win.

2. Cold Calling: The Gatekeeper Problem

If you call a prospect today, their phone likely marks you as “spam risk” automatically. Cognism’s State of Cold Calling Report shows success rates plummeted to 2.3% in 2025. That’s a 50% drop in one year.

3. LinkedIn and Written Outbound: Drowning in Noise

Hunter.io’s 2025 report found that nearly 96% of cold emails go unanswered. Buyers are tuning out the noise. AI email filtering is also getting better and more powerful.

4. The Death of Unqualified Prospecting

Let’s be clear: prospecting isn’t dead. Unqualified prospecting is dead. As Salesloft’s CRO Mark Niemiec noted, AE-generated pipeline converts 3-4x better than traditional SDR motions because the signal is higher.

The Rise of the Human Signal

AI breaks channels because it floods them with volume. Every channel built on volume eventually gets drowned out.

But there’s one thing AI cannot fake at scale: human social proof.

Real usage. Real relationships. Real communities. Real referrals. In the AI era, the most valuable signal isn’t a click or an impression. It’s trust.

As Elena Verna writes, free users with high NPS become your loudest marketers. They share screenshots, evangelize, and generate content. “This is the kind of marketing money can’t buy,” she notes, “and honestly, the kind you could never afford to buy even if you tried.”

The BCG B2B SaaS Growth Report confirms this: best-in-class SaaS companies rely 2-3x more on customer advocacy than their peers.

Why Growth Loops Are Replacing Channels

Traditional acquisition is a funnel. You pour money in the top, and leads trickle out the bottom. The problem? Funnels degrade. Each dollar is spent once and gone. Every month resets to zero.

Growth loops compound. The output of one cycle becomes the input of the next. Users create value that attracts more users.

Many SaaS companies now use referral infrastructure (including tools like Cello) to build these loops directly into their product experience.

Here’s why loops survive the AI flood:

Trust barrier: Loops rely on human-to-human transfer of trust. AI can’t fake a peer recommendation.
Compound interest: Funnels reset to zero. Loops compound. Each cohort helps acquire the next, lowering your marginal CAC over time.
Data moats: Alloy’s research shows integration users are 58% less likely to churn. That’s a switching cost generic AI wrappers can’t replicate.
Ownership: You own the loop. When Google changes its algorithm, your referral program doesn’t break.
AI acceleration: AI can optimize the loop, but it can’t replace the human trust at its core.

Understanding why loops work is easy; building them is hard. We can’t build dynamic systems with static tools. We need a blueprint. This brings us to The Growth OS Map.

The Growth OS Map

To survive, we have to stop thinking in silos and start building an Operating System.

In the old “Funnel View,” Marketing generates leads, Sales closes them, and Success supports them. Data is handed off like a baton. If the baton drops, the customer is lost.

In the “Growth OS View,” these functions are nodes in a network. A happy customer (Success output) immediately becomes a referral source (Marketing input). Sales objections immediately fuel Product updates.

Think of it as five integrated layers:

Layer 1: Signal Over Volume

The foundation is replacing “spray and pray” with signal-based intent. Instead of blasting 10,000 prospects, you identify the 100 showing genuine signs of life:

Product usage patterns
Partner signals
Social engagement
Integration triggers

Signal replaces activity. Quality replaces quantity.

Layer 2: GTM Engineering

This isn’t about hiring more software engineers. It’s about engineering your go-to-market process. We need to break down the walls between Marketing, Sales, and Product. Instead of rigid handoffs, we move to AI-native pods where automation handles the research and data entry, while humans focus on what they’re actually good at: relationships, narrative, and judgment.

Layer 3: Defensible Channels

This is where the loops live. Instead of renting attention, you build compounding mechanisms: viral loops for referrals, content loops for owned media, and integration loops for ecosystem reach. These aren’t campaigns; they are engines that run 24/7.

Layer 4: Retention and Advocacy

Here, we turn satisfaction into fuel. When users become advocates, they feed directly back into the acquisition layer. The math is simple: referred users spend 25% more and churn 18% less, as found in a Wharton study.

Layer 5: Reinvestment

The final layer turns linear growth into exponential compounding. We stop treating outputs (revenue, users, data) as the finish line and start treating them as fuel for the next cycle. Every output becomes an input somewhere else.

• Revenue → Integrations: Profit isn’t just cash; it’s capital used to build new entry points and systematically reduce friction.

• Adoption → Community: A growing user base transforms into a support network that drives its own usage and creates organic advocates.

• Insights → Product Triggers: Usage data doesn’t sit in a dashboard; it converts into automated nudges that boost activation and repeat use.

• Momentum → New Loops: Early traction provides the leverage to launch additional growth loops, ensuring growth multiplies rather than adds.

The Growth Loop Playbook

Not all loops are created equal. We focus on the seven high-leverage loops that build defensible moats.

1. Viral Loop - User Referrals

This loop operationalizes the scarcest resource in the AI era: human trust. It transforms passive user satisfaction into active advocacy by embedding sharing moments directly into the product experience. This mechanism can be fueled with two-sided incentivization to reward both the advocate and the new user. In a landscape saturated with synthetic noise, a personal recommendation provides the high-fidelity signal that buyers are desperate for.

Here’s an example of this: Brex’s referral program.

2. Viral Loop - Partners & Affiliates

While user referrals are organic, partner loops are structured. This scales trust through “nodes of influence”: agencies, consultants, or influencers who already own the relationship with your buyer.

Here’s an example of this: Shopify’s partner/affiliate program.

3. Content Loop (Owned Media)

Newsletters, podcasts, and events create algorithm-proof distribution. Stop relying on renting eyes from platforms and start owning the audience.

Here’s an example of this: Hubspot’s content loop.

4. Community Loop (UGC)

Communities are the new SEO. They resist algorithm volatility by combining User Generated Content (UGC) with social proof.

Here’s an example of this: Tally’s user-generated content loop.

5. Integration Loop - Ecosystem Distribution

Each integration opens a new audience. High-intent users discover you through the tools they already use.

Here’s an example of this: Zapier’s integration growth loop.

6. Product-led Loops

Usage triggers are high-signal GTM inputs. This loop focuses on Casual Contact (CCL). Simply using the product spreads the product.

Here’s an example of this: Calendly’s casual contact loop.

7. Retention Loop

Retention isn’t just about preventing churn; it’s about deep engagement. If you can’t keep them, you can’t grow.

Here’s an example of this: Slack’s retention loop.

8. Onboarding Loop

Activation is the #1 predictor of LTV. This loop bridges the gap between signup and the “Aha!” moment.

Here’s an example of this: Canva’s onboarding loop.

Diagnosing Product-Channel Fit Collapse

You’re experiencing Product-Channel Fit Collapse if:

Your best channels stop producing results even when spent increases.
CAC payback stretches beyond 24-36 months.
Reply rates decline across every outbound motion.
Performance marketing spend becomes unpredictable.
SEO traffic stagnates despite maintaining rankings.
You rely on platforms where visibility can shift overnight.

If three or more of these describe your situation, it’s time for a fundamental GTM rebuild.

How to Rebuild Your GTM Before It Fails

Replace Volume with Signal Kill spray-and-pray. Rebuild around intent, usage, and ecosystem cues. Instead of emailing 10,000 prospects, identify the 100 showing genuine buying signals.
Replace Channels with Loops Shift investment from algorithm-dependent channels to product-embedded loops. Audit your acquisition mix: are you renting or owning?
Replace Silos with a Growth OS Move to a cross-functional GTM engineering pod (RevOps + Growth + Product + AI). Focus on conversion, retention, expansion, and referrals.
Update Your Metrics Dashboard Stop obsessing over “Leads” and “MQLs.” Shift focus to Loop Metrics:
- Payback Period by Channel
- Viral Coefficient
- Network Saturation
- Signal Conversion

The Future Belongs to Teams Who Own Their Distribution

Let’s be clear about what’s happening. Products are easier to build than ever. Channels are collapsing faster than ever.

A defensible asset left is human-earned trust encoded into growth loops.

Users who love your product and tell others. Community members who create content and answer questions. Partners who integrate and co-market. These relationships compound. They can’t be bought. They can’t be automated. They can’t be disrupted by the next algorithm change.

Distribution is no longer a channel problem. It’s an operating system problem.

The companies that thrive will be the ones that are most connected. They’ll turn their customers into advocates, their partners into distribution channels, and their content into media assets. They’ll stop renting their growth and start owning it through loops.

Start now. The easiest loop to activate immediately is the one sitting in your existing happy customer base. Don’t rent attention; unlock the advocacy you already own.

Start with one loop. Make it work. Then add another. The future belongs to teams who own their distribution. Build your Growth OS before someone else builds it around you.

Tag @GTMnow so we can see your takeaways and help amplify them.

Intercom just raised $250M in debt to push its vision of the Customer Agent. Its AI support product Fin already resolves millions of issues weekly, and the next step is bigger: agents that don’t just answer questions, but sell, advise, and manage the full customer relationship. Hear more about Fin from Intercom President Archana Agrawal in our recent GTMnow podcast episode here.

Replit raised $400M at a $9B valuation to push its vision of AI-powered app building. The pitch is simple: turn ideas into working apps without writing code. If that promise holds, the funnel of who can build software just got much wider.

GTM 181: AI at the Edge: How Armada is Taking Compute Everywhere the Cloud Can’t Go | Dan Wright (CEO of Armada)

For the full thing, listen on Apple, Spotify, YouTube or wherever you get your podcasts by searching “The GTMnow Podcast.”

TestBox - announced its Series A alongside three new tools: a synthetic Data Generator, Surface Editor for personalized demos, and Sandbox Copilot, an AI sales engineer. The focus is clear: help sellers move from static demos to interactive proof.

Anything - introduced “design-first vibe coding,” letting builders import Figma designs and go straight into building. It’s an interesting step toward collapsing the gap between design and development.

Esper - raised new funding led by 8VC and Cota Capital following a year of strong growth and leadership expansion. The company focuses on modern device infrastructure for fleets of dedicated Android devices across retail, logistics, and enterprise environments.

Founding Marketing & Growth Lead at Village (Hybrid - Los Angeles, CA)
Community Partnerships Lead at Village (Hybrid - Los Angeles, CA)
Founding Enterprise AE (New Business & Outbound Focus) at Testbox (Remote - San Francisco, CA)
Revenue Operations Manager at Vanta (Remote - US)
Business Development Representative at Northbeam (Remote - US)
Strategic account executive at Writer
1. Central (Hybrid - Chicago, IL; Austin, TX; Dallas, TX; Indianapolis, IN; Minneapolis, MN)
2. East (Hybrid - New York City, NY; Boston, MA; Hartford, CT; Jersey City, NJ; New Haven, CT; Newark, NJ; Philadelphia, PA; Stamford, CT)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

Spryng (for marketers): March 24–25, 2026 (Austin, TX)
MicroConf 2026: April 12–14, 2026 (Portland, OR)
GTMfund Dinner: April 14, 2026 (Austin, TX)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
GTMfund Dinner: May 14, 2026 (San Francisco, CA)
GTMfund Dinner: June 9, 2026 (London, UK)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow community (founder, operator, investor) love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Stefan Bader and Sophie Buonassisi (not AI!).

GTM: The cloud covers 30% of the world. Armada is building the other 70%

GTMnow — Wed, 11 Mar 2026 17:13:43 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by:

If you run go-to-market, you already know the problem: your data lives everywhere. Spreadsheets, CRMs, sales calls, ad platforms… yet you’re still guessing what to do next.

Learn more at hockeystack.com

Who we sat down with

Dan Wright (Co-founder & CEO of Armada) joins GTMnow to unpack what it actually takes to bring AI infrastructure to the places the cloud was never built to reach.

The cloud covers about 30% of the world. The other 70% (think: oil rigs, the Arctic tundra, military ships, remote mines) is where some of the most critical decisions happen, making latency a life-or-death and billions of dollars difference. Armada is building the infrastructure for that part of the world: modular, ruggedized AI data centers that go to the data, instead of the other way around.

From the first offshore edge computing deployment with the US Navy, to cutting avalanche response times in Alaska from over a day to real time, to sovereign AI installations in Saudi Arabia with Aramco and Microsoft, Armada is redefining what operating at the edge even means.

Discussed in this episode

Why cloud infrastructure was built for a pre-AI world and what that gap costs
How Starlink turned every remote location into a potential AI cluster
What “distributed intelligence” means and why it’s the founding principle behind Armada
The global race for AI sovereignty and why modular compute is the linchpin
How Armada goes to market when a product demo involves shipping a 40-foot container to a desert (yes, really)
Why customer champions are better than any sales rep
The Microsoft partnership and how Armada extends Azure to places Azure could never go on its own
Category creation lessons from building a company before the market had a name for the industry
What’s next: SpaceX, sovereign AI, and why Dan thinks humans are on the moon in two years (yes, REALLY)

Episode highlights

00:00 – Why the cloud only reaches 30% of the world and what Armada actually builds

03:11 – From Saudi Arabia to the Arctic: recent deployments that redefine the edge

04:12 – AI’s physics problem: why distance from data breaks everything

05:31 – How Starlink turned every remote location into a potential AI cluster

07:14 – What distributed intelligence means and why it’s Armada’s founding principle

08:14 – The global AI race: sovereign compute as a national security strategy

09:00 – The Genesis Mission, the White House, and Davos: why sovereign AI is mainstream

12:49 – GTM for a hardware company when the demo involves shipping a 40-foot container

15:18 – Why Armada’s customers do the selling for them (and do a better job than some reps)

17:35 – The Microsoft partnership and extending Azure to the places it can’t reach on its own

18:51 – Category creation: the lesson Dan learned about specificity and business value

21:02 – What’s next for Armada, SpaceX, and why Dan thinks we’re two years from the moon

25:34 – The founder advice Dan wishes more people took seriously

View the Full Transcript

Key takeaways

1. The cloud was built for a world before AI.
Every major cloud provider built their infrastructure 25 years ago, before large models existed, before sensors were on everything, before data lived on oil rigs or in the Arctic. The world has changed but the infrastructure has not. That mismatch is a massive market opportunity.

2. Connectivity unlocked the edge, but without infrastructure it’s useless.
Starlink changed the game for connectivity. But without compute, all you have is faster access to a problem you still can’t solve. Every time Starlink goes live in a new country, Armada is the first mover on infrastructure. Connectivity + compute = the edge finally works.

3. Sovereign AI is here and the US and China are in a global AI race.
We’re in a global AI race. And infrastructure is the battleground. For companies like Aramco and the US Navy, data sovereignty means the data cannot leave the site - by law, by security protocol, by necessity. You can’t send classified sensor data from an oil rig or a Navy ship to a data center thousands of miles away and wait for a response. You need compute where the data lives (that’s what Armada builds).

4. Category creation is about problems, not technology.
Nobody wakes up wanting to buy infrastructure. Dan’s approach from day one was to get specific about the problems (avalanche response in Alaska, split-second decisions on a battlefield, catastrophic failure prevention on an oil rig) and let the technology follow.

5. Write your manifesto before you do anything else.
Before a single hire, before an office, before raising a dollar. Armada had a manifesto - this is why we’re building, here’s where the world is going, here’s why it matters. Dan calls it an unfair advantage. When you can articulate exactly why your company needs to exist, recruiting gets easier, fundraising gets easier, partnerships get easier. You’re not selling a product. You’re building a movement. A 10-slide pitch deck might get you to Seed, but it won’t build a generational company.

Follow Dan Wright

LinkedIn: https://www.linkedin.com/in/wrightdh
X: https://x.com/danwrightSF
Armada’s LinkedIn: https://www.linkedin.com/company/armadaai
Armada’s website: https://www.armada.ai

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast tells the stories of how the top 1% of founders, operators, and investors build, scale and invest.

Each week, we uncover the pivotal moments, bold decisions, and go-to-market strategies that turn ideas into breakout companies.

GTMnow is the media brand of GTMfund - we are an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

The "SAASpocalypse”

GTMnow — Fri, 06 Mar 2026 17:59:11 GMT

This is an important edition addressing sentiments around a looming “SAASpocalypse.” It was first published yesterday on X and is an article written by GTMfund’s General Partner, Max Altschuler. If you’re wondering about AI’s impact on software as a whole, this is a deep dive written in Max’s first person perspective below. Let’s get into it.

In the 5+ years we’ve been building GTMfund, I’m not sure there’s been a more narrative-dense month. At a minimum, not since ChatGPT launched in late 2022.

B2B SaaS is dead. Public market software reset. OpenClaw. AI is eating the world.

I wrote this for anyone watching these headlines and feeling uncertain about what they mean. It’s our thoughts on the “SaaSpocalypse,” public markets, and what that means for the startup ecosystem.

Let’s start with the death of SaaS. It’s important to split this conversation into two distinct and important points:

SaaS is dead because customers will vibe code their own solutions.
Public markets are revaluing software multiples because of disruption on the horizon.

These are the two primary drivers of panic in the market. We have a lot less concern about the first point.

SAASpocalypse reason #1: customers will vibe code their own solutions.

AI has created a reality where the incremental cost and difficulty to produce software are declining rapidly. It’s never been easier or faster to code. For the first time in history, non-technical people are writing code and creating products. Even if most of those are prototypes today, the trend is obvious and inevitable – the barriers to entry for creating a usable piece of software are declining to zero.

Naturally, that begs the question: if the cost of creating net-new software continues to decline, won’t customers create their own custom version of the applications they need instead of buying from a vendor? This is the thesis behind vibe coding killing modern SaaS.

The problem with that position is it misses the primary reason most companies buy software. Historically, companies haven’t purchased software because there was no pathway for them to create the tools internally. That was a constraint, but it was never the core driver of purchasing behavior.

Let’s take DocuSign as an example. For an enterprise buyer, the decision isn’t about whether their team could technically build an e-signature tool. It’s about trust, compliance, and legal enforceability. Contracts need to hold up in court. And for the cost of a DocuSign seat (or per-contract pricing if they shift to a usage-based pricing model), that tradeoff is obvious. The expenditure is justified because every signed agreement is legally binding and defensible. The workflows and additional functionality are a bonus, but no serious company is going to risk the legality of its contracts just to save $10 per seat per month and vibe code a replacement.

A similar principle applies to more complex or foundational systems, like CRMs or ERPs. The constraint isn’t technical feasibility. In many cases, a company could build an internal version. The real question is whether it’s rational to do so.

The cost to host, manage, secure, patch, and continuously maintain a platform of that complexity far outweighs the perceived savings of building in-house. Those resources are better allocated toward core product innovation and differentiated capabilities that actually move the business forward. Any OpEx savings are quickly eroded by the long-term operational burden of maintaining and evolving that system.

So, what does vibe coding kill? It kills point solutions. It kills surface-layer applications where a highly customized internal tool may deliver more marginal value than a generic off-the-shelf product. But the thing is, many of those companies were structurally fragile even before AI. They were going to struggle either way – AI just makes the pain of that realization more acute. Historically, some point solution companies could go on to raise Series A, Series B, and sometimes beyond. They would show early traction and promise. But eventually, they would hit a ceiling. AI has accelerated the timeline to that ceiling, but it’s always existed. This is why we have never oriented our investment strategy around point-solution application companies, and never will.

Finally, in the technology ecosystem, we have a tendency to look inward. We live in an echo chamber to an extent. Just because an AI-native Series B company is building a swarm of GTM agents on top of their open-source database vs. buying a CRM doesn’t mean Avis or Coca-Cola is going to do the same. Or any non-tech company. They don’t have the resources, expertise, or interest to divert that level of mindshare away from their core product or mission. It’s just not worth the trade-off. They’ll become AI native, but they’ll do that by purchasing AI-native software – just like they always have. After all, global business software spend is at its highest level in history and accelerating.

This brings us to SAASpocalypse reason #2: public markets are revaluing software company durability and future cash flows.

This challenge is real. It’s more visible in the public markets, but the late-stage and growth-stage private markets are undergoing the same shift. Historically, software companies get better multiples than the average company in the public market because of the reliability of their cash flows. Companies invest a lot in R&D upfront, but once they reach a certain state of maturity, they sell recurring annual contracts at attractive margins (often 80%+) with accounts that expand organically (100%+ NRR). To investors, these businesses were treated almost like annuities. Even without significant net-new growth, a traditional software company could generate predictable internal expansion, low churn, and high-margin revenue.

AI is changing the equation, and it’s changing the math on two fronts:

Increased competition (and therefore higher potential churn).
Value squeeze on the middle layer of the technology stack (will expand on this below).

First, the core concern with the declining cost and difficulty of producing usable software is not that customers will vibe code their own enterprise platforms, as outlined earlier. The more material impact is on competitive dynamics. In the cloud era, engineering talent was a bottleneck. Companies could build defensibility by getting to market first, hiring strong engineering teams, and compounding product advantage over time. That bottleneck no longer exists. A small, highly capable team can now build in weeks what previously took years. The market will get flooded with competitive solutions (it already is), and customers will have more choice than ever for which vendor to go with.

On top of this, one of the pillars of software defensibility in the cloud era was switching costs. If you wanted to rip and replace an existing vendor, you had to convince the customer that it would be worth the data transformation, retraining their entire staff, managing change, etc. Well, today, a swarm of AI agents can do the data transformation. You don’t need to relearn a new technology platform if the UI of your application can be reduced to a chat interface and some simple workflows. Declining switching costs directly challenge the defensibility of application-layer SaaS. The markets are responding to this. If someone can’t reliably forecast that Customer X is going to pay a company $100k+ a year for the next decade, well, they’re going to discount those future cash flows more than they previously did. That will lower the company’s market multiple.

The second phenomenon in the revaluation story is the “Middle Squeeze.” There are lots of great reads on this topic specifically, but we’ll reference a mental model shared by David Ondrej (full article in the Quality Reads section).

In an AI-native reality, where value accrues is shifting. Traditional application SaaS existed a lot in the middle layer. Sure, there were some actions you could execute from your SaaS platform (top layer), but they weren’t true agentic outcomes. Middle-layer platforms just made the humans using software on the other end more efficient, but the humans were still the top layer.

AI is pushing value accrual to either end of the equation – the top layer (agents executing workflows and driving outcomes autonomously) and the bottom layer (data that powers these agents). If you’re a Middle-Layer public market software company, you might be in trouble. Agents are coming for your workflow, and they’ll be able to execute it faster and at scale. Over time, they’ll likely be able to execute it better than humans.

If you are a Middle-Layer and Bottom-Layer company (like Salesforce or HubSpot), the public markets will value you more for your bottom-layer capabilities. If you become a database company in the future, suddenly you may not be able to charge the same price per seat, expect the same growth in seats over time, or both. You either need to build a top-layer solution (like Agentforce) that the market loves or risk being relegated to the bottom layer for good. It’s no wonder investors are revaluing these businesses, as this is a tectonic shift in the ecosystem.

Now, what does that mean for us as investors in early-stage B2B technology? Or what does that mean if you’re an operator at early-stage software companies? Not as much as you’d think. Software isn’t dead. Where the value accrues is shifting and shifting rapidly. This means a larger-than-normal portion of incumbents are at risk. And because of how fast AI moves, that risk is building faster than ever before.

However, the underlying principles of value remain: B2B technology spend is growing faster than ever before. AI will capture portions of traditional software budgets but also expand into services, agency, and labor spend. The opportunity is larger than ever; you just need to be thoughtful and deliberate about where you invest and where the value will accrue over time.

Traditional SaaS might be dying. B2B technology is not – it’s alive, well, and growing. Our job is to invest in the future of B2B technology, not the software platforms of yesterday.

Tag @GTMnow so we can see your takeaways and help amplify them.

Who is actually accelerating in B2B right now. Software stocks are way, way down in 2026, but that doesn’t mean growth is down for many of the best.

In defense of vertical software. Right now everyone in this industry is asking some version of the same question: in a world where Anthropic, OpenAI, and Google are pouring hundreds of billions of dollars into building the most powerful general-purpose AI systems in the world, why does vertical software need to exist?

VC episode: 116 Quarters on Quota: The Sales Metrics That Actually Matter with Bill Binch, Operating Partner at Battery Ventures

GTM 181: Lessons from Figma, Dropbox, and Building Iconic Brands in the Age of AI with Figma’s CMO, Sheila Vashee

For the full thing, listen on Apple, Spotify, YouTube or wherever you get your podcasts by searching “The GTMnow Podcast.”

Nominal - Hardware testing startup hits $1B valuation, raises $155M in 10 months. Nominal on Thursday announced a fresh $80 million Series B extension round at a $1 billion valuation.

Nooks - just launched AI Sequencing, extending beyond its dialer into full multi-channel outbound. Lists, research, email drafting, signals, coaching — all inside what they’re calling the Agent Workspace.

VP of Sales at Spekit (Hybrid - Denver, CO)
Strategic Account Executive (Central) at Writer (Remote - US)
AI Native Growth Marketer at Wynter (Remote)
Growth Development Representative at Northbeam (Hybrid - San Francisco / Los Angeles)
Sales Manager, Vanta for Government at Vanta (Remote - US)
Sales Manager at Owner.com (Remote - Toronto, ON)
AI Engineer, Reinforcement Learning at Oureon (Austin, TX)
Software Engineer, Autonomy at Oureon (Austin, TX)
Backend Engineer at Oureon (Austin, TX)
Software Engineer, Forward Deployed at Oureon (Austin, TX)
Frontend Engineer at Oureon (Austin, TX)

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

MicroConf 2026: April 12–14, 2026 (Portland, OR)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
Demand & Expand: May 19-20, 2026 (San Francisco, CA)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow community (founder, operator, investor) love to close it out - we appreciate you.

Subscribe now

This newsletter was written and edited by Max Altschuler and the GTMfund team (not AI!).

GTM: “Make good $h!t.” Figma’s CMO on brand in 2026

GTMnow — Wed, 04 Mar 2026 18:05:09 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by:

If you run go-to-market, you already know the problem: your data lives everywhere. Spreadsheets, CRMs, sales calls, ad platforms… yet you’re still guessing what to do next.

Learn more at hockeystack.com

We’ve got an incredible deal exclusive to GTMnow listeners.

It’s for Granola, the AI notepad for people in back-to-back meetings.

We’re Granola users at GTMfund, and it has changed the way we work.

Granola takes meeting notes for you without any intrusive bots joining your calls.
During or after the call, you can chat with your notes.
Ask Granola to pull out action items, help you negotiate, write a follow-up email … or even coach you using Recipes, which are pre-made prompts.
It’s actually the same technology we use to create the notes for this very podcast. Once you try it on a first meeting, it’s hard to go without.

Head to granola.ai/gtmfund and get three months free with the code GTMFUND.

Who we sat down with

Sheila Vashee (CMO of @Figma) joins GTMnow to share how she thinks about brand building across every stage. From selling brownies at age eight, to second marketing hire at Dropbox scaling to $1B+ in annualized revenue, to now leading marketing at one of the most beloved software brands in the world, Sheila has seen it all.

In this conversation, you’ll learn what brand actually means (hint: it’s not your logo), how PLG companies make the leap to enterprise, why being obsessively close to your customers is a compounding advantage, and how AI is reshaping the marketing playbook without replacing the human craft that sets great brands apart.

Discussed in this episode

Why brand is the sum total of every experience (not just pretty visuals)
What being the second marketing hire at Dropbox through $1B+ in revenue taught her about early-stage strategy
How Dropbox’s Space Race campaign blended marketing into company strategy
The PLG-to-enterprise equation: what Figma did early that Dropbox waited too long on
Why building the enterprise team early is both an operational and an optical investment
How Figma ingests customer feedback at scale using quantitative tools + live user research
The role of AI at Figma: enabling human creativity, not replacing it
What the shift from SEO to GEO means for marketing teams
Why social and third-party validation are back as core growth levers
How consumer marketing experience shaped her approach to measuring brand ROI
The mentors that shaped her career and the question that changed her trajectory
Her one piece of advice for anyone building a brand right now: “make good $h!t”

Episode highlights

00:00 – The brownie stand: why brand building started at age eight

03:55 – How Sheila defines brand: it’s what people say when you’re not in the room

05:08 – Joining Dropbox as the second marketing hire

06:07 – Space Race: the campaign that defined Dropbox’s early strategy

06:56 – What consumer marketing taught her about driving revenue

08:29 – Measuring brand ROI through match market testing

09:15 – How Figma thinks about brand building at a macro level

11:21 – The PLG-to-enterprise equation: what Figma did early that Dropbox waited too long on

15:02 – Why building the enterprise team is both operational and optical

17:12 – How Figma ingests customer feedback at scale

18:26 – AI at Figma: enabling human creativity, not replacing it

21:13 – What the venture side taught her about staying sharp as an operator

22:33 – Why the shift from SEO to GEO is inevitable and what to do about it

23:24 – Why Reddit is back and social is a core growth lever

24:51 – The mentors that changed her trajectory

28:59 – One piece of advice: make good $h!t

View the Full Transcript

Key Takeaways

1. Brand is the sum total of every experience (not a logo or a campaign).
It’s what people say about you when you’re not in the room. It’s how your support team talks to customers, how your salespeople engage buyers, and the experience your product delivers. Every person at the company is a node of brand.

2. The PLG-to-enterprise leap requires an intentional, early investment.
Figma built an enterprise team far earlier than most PLG companies do and it paid off. It’s the reason 95% of the Fortune 500 now uses Figma. But optics matter more than you’d think: enterprise customers need to see you’re invested in their success before they’ll bet on you.

3. Customer obsession is the new operating model.
Figma’s founders flew to Nigeria to meet early users. Their product support team feeds insights directly back to product. They have 200+ Friends of Figma chapters running worldwide. And their user research sessions are open for anyone at the company to watch live. That closeness to the customer is embedded in the Figma DNA, and it shows up in the product.

4. Measuring brand impact requires creativity.
Sheila ran match market tests at a previous company by fully treating one city with TV, billboards, and brand spend while keeping comparable cities at steady state. The ROI was clear. For brand skeptics who say you can’t quantify it: design the experiment, isolate the variable, and let the data speak.

5. AI enables human creativity. Human taste will always win.
Figma sees AI as a massive enabler. Tools like Figma Make let you generate 20 prototypes in 20 different directions. But the judgment on which path to take, the craft that brings it to life, that’s still human. Sheila’s push to the marketing community: don’t fall into the AI slop trap. MAKE GOOD $h!t!

Follow Sheila Vashee

LinkedIn: https://www.linkedin.com/in/sheilavashee
Figma’s LinkedIn: https://www.linkedin.com/company/figma
Figma website: https://www.figma.com/

Follow Sophie Buonassisi (Host)

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast tells the stories of how the top 1% of founders, operators, and investors build, scale and invest.

Each week, we uncover the pivotal moments, bold decisions, and go-to-market strategies that turn ideas into breakout companies.

GTMnow is the media brand of GTMfund - we are an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

VC: 116 Quarters on Quota: The Sales Metrics That Actually Matter

GTMnow — Tue, 03 Mar 2026 20:39:03 GMT

Listen on YouTube

Listen on Apple

Listen on Spotify

Brought to you by: AngelList

They’ve scaled with us across funds and into the future.

If your fund is growing in size or complexity, check them out at https://angellist.com/gtmfund.

Who we sat down with

Bill Binch (Operating Partner at Battery Ventures) joins GTMnow to share the operational frameworks he’s built across 116 quarters on quota, and what actually changes when you move from driving revenue to advising an entire portfolio.

Before Battery, Bill was employee #16 at Marketo, where he led sales from zero revenue through IPO and a Vista Equity acquisition. He then served as CRO at Pendo, helping scale ARR to nearly $100M in only three years. His career spans some of the most defining sales organizations in enterprise software, from Oracle to PeopleSoft.

Discussed in this episode

Why every forecast conversation should start with “my quota is” not “my forecast is”
The three metrics most companies aren’t tracking (and should be)
The “mojo metric”: 6 daily inputs that tell you whether pipeline is gaining or losing ground
Why the 5-quarter look-back is the most important slide in any board deck
What the transition from CRO to operating partner actually looks like
Why Battery doesn’t run a one-size-fits-all playbook across its portfolio
The accidental sales experiment at Outreach that proved the power of in-person teams
How the best sales and marketing leaders present as one motion in board meetings
What Bill looks for when evaluating whether a company has the operational rigor to scale
His advice for VPs, CROs, and CMOs who want to eventually move into an operating partner role

Episode highlights

00:00 – Bill’s career path: Oracle, Marketo employee #16, Pendo CRO, and choosing Arizona over tech hubs

15:01 – Battery’s approach: why there’s no one-size-fits-all playbook

19:56 – Outreach’s accidental sales experiment: remote vs. in-person results

21:23 – What nobody tells you about the transition from CRO to operating partner

23:01 – Gold watch and slippers? Why operating partner is not a retirement plan

27:15 – Three audiences: prospective investments, portfolio companies, and Battery internal

31:42 – Investing in GTM tech: why Battery went 7 years between bets

34:44 – The Phil Fernandez warning: “your hand is on the wheel, now you’re a passenger”

36:44 – Why every forecast should start with “my quota is” not “my forecast is”

38:33 – The metric most companies miss: plan vs. planned quota vs. actual quota deployed

40:05 – The mojo metric: 6 daily pipeline inputs that tell you everything

43:16 – Helping technical founders sell: the one-eyed man in the land of the blind

48:04 – Advice for operators who want to move into an operating partner role

51:03 – The sales and marketing dance: what real GTM alignment sounds like in a board room

View the Full Transcript

Key takeaways

1. Every forecast conversation should anchor to quota, not forecast.
Bill requires reps to start with three numbers in order: my quota is, my forecast is, my closed won is. Most reps default to leading with forecast, which lets the conversation drift from the actual target. When you anchor to quota first, the gaps become immediately visible, and you know exactly how much urgency is required. Sales leaders get paid on quota, not forecast. Language used should reflect that.

2. Most companies don’t track whether they’ve actually deployed enough quota to hit plan (and that’s a massive mistake).
Three metrics, tracked monthly: company plan, planned quota deployed, and actual quota deployed. Bill gets called all the time by leaders in Battery’s portfolio saying “we don’t have enough pipeline.” The first place he looks? Quota deployment, because more often than not the real problem is they’re running at 70% of planned capacity but expecting 100% of plan. That math is never going to work, and reveals that hiring is your problem, not quota.

3. Pipeline health is a daily exercise.
The “mojo metric” tracks six daily inputs: three that add to pipeline (new deals, expansions, pull forwards) and three that subtract (killed, shrunk, pushed). Net them out every night. Three straight days of net negative and you’re sounding the alarms. Foolproof way to catch slow leaks overnight instead of once the quarter is already gone.

4. The transition from CRO to operating partner is more of an identity shift than people expect.
After 29 years with a clear scorecard, Bill moved into a role where outcomes might be 5-10 years out. Measuring impact is entirely different in venture. Be prepared to solve a completely different set of problems requiring a lot more context switching than when success metrics were pre-defined.

5. The path from operator to operating partner starts years before you’re ready.
If you’re working for a venture or PE-backed company, build relationships with those teams now. Show up at industry events, chat with the team. It’s a sales pipeline nurture, just for your career.

Follow Bill Binch

LinkedIn: https://www.linkedin.com/in/bill-binch-302a4a2
Battery Ventures’ LinkedIn: https://www.linkedin.com/company/battery-ventures
Battery Ventures’ website: https://www.battery.com

Follow Max Altschuler and Paul Irving

Max Altschuler on LinkedIn: https://www.linkedin.com/in/maxaltschuler
Max Altschuler on X (Twitter): https://x.com/HackItMax
Paul Irving on LinkedIn: https://www.linkedin.com/in/paulsirving
Paul Irving on X: https://x.com/PaulGTM
Newsletter:

Follow GTMnow

LinkedIn: https://www.linkedin.com/company/gtmnow
X / Twitter: https://x.com/GTMnow_
YouTube: https://www.youtube.com/@GTM_now
Podcast directories: https://gtmnow.com/tag/podcast

Follow GTMfund

Newsletter: You’re in it :) If this was forwarded to you, you can subscribe here. GTMnow is our main brand for content, but we also have two around our investments:
LinkedIn: https://www.linkedin.com/company/gtmfund
X: https://x.com/gtmfund

The GTMnow Podcast - subscribe to support

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify

The GTMnow Podcast tells the stories of how the top 1% of founders, operators, and investors build, scale and invest.

Each week, we uncover the pivotal moments, bold decisions, and go-to-market strategies that turn ideas into breakout companies.

GTMnow is the media brand of GTMfund - we are an early-stage venture firm made up of 350+ go-to-market executives from the fastest-growing companies.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

Claude Code for GTM Teams

GTMnow — Fri, 27 Feb 2026 18:26:02 GMT

Last week, we hosted a private GTMfund Claude Code workshop for founders alongside AI expert, Jordan Crawford. Across that session and in conversations with hundreds of founders and GTM leaders each month, this is something we’re hearing:

Interestingly, the biggest blocker around AI for people isn’t always technical, but rather involves the imagination - knowing what is actually possible.

We wrote this edition to help unblock the imagination block, specifically for people who run go-to-market.

Why specifically Claude Code? Dan Shipper of Every offered a brilliant analogy: “The cloud app is like a hotel room, clean, set up for you, but you start fresh each time. Claude Code is like having your own apartment with AI in it.”

This edition gives you the apartment tour. Setup in 5 minutes, then 30+ use cases across every GTM function, with the exact prompts to get started.

First, What is Claude Code (and why should you care)?

Claude Code is Claude running in your terminal, that text-based interface your computer already has. Think of it as an execution layer, rather than a coding tool.

Where the regular Claude chat is a brilliant advisor you talk to, Claude Code can actually do things on your machine:

Access all your local files: no uploading, no size limits. Drop 500 call transcripts into a folder and tell it to analyze them.
Run for hours on complex tasks. The cloud app times out. Claude Code keeps going.
Execute multi-step workflows: it plans, acts, checks its work, and iterates. Not a chatbot, a true agent.
Install and use tools: it can set up scripts, connect to APIs, browse the web, and create files directly on your machine.
Work across every file type: PDFs, CSVs, audio files, images, and spreadsheets. If it’s on your computer, Claude Code can work with it.

Boris Cherny, the creator of Claude Code, originally built it as an internal tool at Anthropic. It wasn’t supposed to be a product. But people started using it for way more than coding - for organizing files, writing status updates, and automating busywork. The use cases just kept expanding.

If you’re already using Claude Cowork (the GUI version that shipped in January), think of Claude Code as Cowork with training wheels removed, faster, more flexible, and with more direct access to your system. Both use the same underlying agent architecture.

Install Claude Code in 5 Minutes

This is the part that scares people. It shouldn’t. Here’s the entire process:

Step 1: Open Your Terminal

On Mac: Press Command (⌘) + Space, type “Terminal,” and hit Return.

On Windows: Press the Windows key + R, type “wt,” and press Enter.

Step 2: Install Claude Code

Mac:

curl -fsSL https://claude.ai/install.sh | bash

Windows:

irm https://claude.ai/install.ps1 | iex

Step 3: Launch It

claude

That’s it. You’ll be prompted to log in with your Anthropic account. Claude Code works with any paid Claude plan ($20/month Pro or $100/month Max for heavier usage).

The Use Cases: A GTM Operator’s Claude Code Playbook

Below is a structured breakdown of real use cases across every GTM function. Each includes the scenario, the exact prompt (or prompt pattern), and where applicable, what the output looks like. We’ve organized them by function so you can jump to what’s most relevant to you.

1. Personal & Admin

The “I can’t believe I used to do this manually” category.

Organize Your Disaster of a Downloads Folder

Every operator’s Downloads folder is a graveyard of random PDFs, screenshots, and files named “final_v3_REAL_final.docx.”

Claude Code can analyze hundreds of files by type, date, and content, then reorganize them into a logical structure automatically.

Prompt: “Analyze my Downloads folder and organize everything into logical subfolders by: file type (Documents, Images, Videos), date, and project or category. Propose a plan before moving anything.”

Sort Invoices for Tax Season

Drop your messy pile of invoices and receipts into a folder. Claude Code reads each file, renames it with a standardized format, and sorts it into the right category.

Prompt: “Read each file in this folder. Rename them to ‘YYYY-MM-DD Vendor - Invoice - ProductOrService.pdf’ and move them into subfolders by quarter (Q1, Q2, Q3, Q4).”

Clear Storage on Your Machine

Instead of manually hunting for what’s eating your disk space, just ask.

Prompt: “Scan my computer for files larger than 500MB. Show file name, location, size, and last accessed date. Sort by size. Help me free up at least 20GB safely.”

Draft and Schedule Social Posts

Feed Claude Code your recent content (blog posts, transcripts, notes), and ask it to generate a week’s worth of LinkedIn posts in your voice.

Prompt: “Read all the markdown files in my /content folder. These are my recent blog posts. Generate 5 LinkedIn posts based on the most interesting insights from these pieces. Match my writing style: direct, analytical, no fluff. Save each as a separate .md file in /social-drafts.”

💡 Pro tip: Pair this with a CLAUDE.md file that contains your voice guidelines and style preferences. Claude Code reads this file automatically at the start of every session, so it always writes in your tone.

Build a Personal CRM from Your Inbox

Export your email contacts or meeting notes into a folder. Claude Code can parse them and build a lightweight relationship tracker.

Prompt: “Analyze all the .eml files in this folder. Extract: sender name, company, date, and key topics discussed. Create a CSV with columns for Name, Company, Last Contact Date, Topics, and Suggested Follow-Up. Flag anyone I haven’t contacted in 30+ days.”

2. Sales

Identify High-Quality Leads from Your Codebase or Product

This is one use case from Lenny’s roundup. If you’re a founder, run Claude Code from your app’s source directory and ask it to analyze what you’re building to suggest ideal pilot customers.

Prompt: “Look at what I’m building in this directory. Based on the product functionality, identify the top 10 companies in [your city/region] that would be ideal pilot customers. Include their likely pain points and a suggested outreach angle for each.”

Scrape Public Data for Prospect Lists

One user in Lenny’s community used Claude Code to scrape GitHub repos to find companies with sensitive data exposure, then pitched them his security product. The result: a prioritized lead list with LinkedIn URLs and a relevance score.

Prompt: “Search GitHub for public repositories that contain [specific indicator relevant to your product, e.g., hardcoded API keys, unencrypted credentials]. For each, extract: repo owner, company name if available, repo URL, and a priority score (1-10) based on severity. Export as CSV.”

Analyze Call Transcripts at Scale

Download your Gong, Chorus, or Fireflies transcripts into a folder. Let Claude Code do the work of a sales ops analyst.

Prompt: “Analyze all call transcript files in this folder. For each, extract: prospect company name, key objections raised, competitor mentions, pricing discussion points, and deal stage signals. Then create a summary report showing: (1) the top 5 most common objections across all calls, (2) which competitors are mentioned most, and (3) which calls have the strongest buying signals.”

Generate Personalized Outreach Sequences

Feed it a prospect list CSV and your product context.

Prompt: “Read prospects.csv in this folder. For each prospect, research their company using the web and draft a 3-email outreach sequence. Email 1: personalized hook based on a recent company event. Email 2: problem-specific value prop. Email 3: breakup email with a resource link. Save all sequences to /outreach-drafts as individual text files.”

Competitive Battlecard Builder

Point Claude Code at competitor websites and your own positioning docs. Get structured battlecards your reps can actually use.

Prompt: “Read our product positioning doc at /docs/positioning.md. Then research [Competitor 1], [Competitor 2], and [Competitor 3] from their websites. For each competitor, create a battlecard with their positioning, key differentiators, common objections they raise against us, our best counter-arguments, and scenarios where we win vs. lose. Save as battlecards.md.”

3. Marketing

The function where Claude Code might deliver the highest ROI per hour invested.

Competitor Ad Intelligence

Use Claude Code to pull and analyze competitor ads from Meta’s Ad Library. No more manual screenshot sessions.

Prompt: “Go to Meta’s Ad Library and search for ads from [Competitor]. Download the top 20 active ads. For each, analyze: messaging angle, CTA, visual style, and target audience signals. Create a competitive ad report with recommendations for how we should differentiate our own ad creative.”

Content Performance Analyzer

Export your blog analytics or newsletter data as a CSV. Claude Code can find patterns you’d miss in a spreadsheet.

Prompt: “Find the CSV file in my Documents folder called blog_performance_2025.csv. Analyze every post and surface patterns: which topics get the highest open rates, what headline formats perform best, what publish day/time correlates with engagement, and what content length hits the sweet spot. Create a visual report with charts.”

SEO/GEO Keyword Research Pipeline

As SEO shifts toward Generative Engine Optimization (GEO), you need to understand what AI search engines surface. Claude Code can build you a research pipeline.

Prompt: “Research the top 50 keywords in [your industry vertical]. For each, check: current SERP ranking difficulty, whether AI overviews appear for that query, estimated search volume, and our current ranking if available from /data/rankings.csv. Export a prioritized keyword matrix ranked by opportunity score.”

Repurpose Long-Form Content into Multi-Channel Assets

Feed it a blog post, webinar transcript, or podcast episode. Get back a full content repurposing kit.

Prompt: “Read /content/webinar_transcript_q4.md. From this single piece, create: (1) a 1,500-word blog post, (2) a Twitter/X thread of 8 tweets, (3) 3 LinkedIn posts with different angles, (4) 5 email subject lines for newsletter distribution, and (5) a one-page executive summary. Save each as a separate file in /repurposed.”

Build a Self-Updating Competitive Tracker

This is a power-user move. Create a script that Claude Code runs periodically to monitor competitor changes.

Prompt: “Create a script that checks these competitor URLs every week: [URL 1], [URL 2], [URL 3]. Track: pricing page changes, new feature announcements, job postings (especially in sales and engineering), and press mentions. Append changes to /competitive-intel/changelog.md with timestamps.”

4. Customer Success

Renewal Risk Analyzer

Export your customer health data or usage metrics. Let Claude Code flag accounts that need attention before it’s too late.

Prompt: “Analyze customer_health.csv in this folder. For each account, flag risk level (Green/Yellow/Red) based on: usage trend (declining = red), support ticket volume (increasing = yellow/red), NPS score (below 7 = red), and days until renewal. Create a prioritized action list for the CS team with specific suggested interventions for each Red account.”

QBR Deck Generator

Instead of spending half a day pulling data into slides, feed Claude Code your account data and let it draft the QBR.

Prompt: “Read the usage data, support tickets, and NPS survey results for [Account Name] from the /accounts folder. Generate a QBR document that includes: executive summary, usage highlights and trends, ROI metrics, support summary, product roadmap alignment, and recommended expansion opportunities. Save as a markdown file.”

Synthesize Customer Feedback at Scale

Use Claude Code to process call transcripts and map feedback against your product assumptions.

Prompt: “Read all files in /customer-calls. For each transcript, extract: feature requests, pain points, praise, and competitor mentions. Then compare these against the assumptions in /docs/product-hypotheses.md. Create a scorecard showing which hypotheses are validated, which are invalidated, and which need more evidence.”

Churn Post-Mortem Analyzer

After a wave of churn, don’t just guess at the reasons. Run the data.

Prompt: “Analyze churned_accounts_q4.csv and all related support transcripts in /churn-data. Identify the top 5 churn drivers by frequency. For each driver, include: number of affected accounts, representative quotes from transcripts, and a suggested preventive action. Create an executive summary for leadership.”

5. Product

Write Job Descriptions and Hiring Plans

Justin Bleuel from Clay shared this use case: feed Claude Code your existing hiring materials and comparable JDs from other companies. Get a complete hiring package.

Prompt: “Read the hiring materials in /hr/pm-hiring and the competitor JDs I’ve saved in /hr/reference-jds. Generate: (1) a complete job description for a Senior Product Manager role, (2) a structured interview plan with 4 rounds, (3) an evaluation rubric with scoring criteria, and (4) a 90-day onboarding plan. Save each as a separate doc.”

Self-Driving Documentation

James Pember coined this term: using Claude Code with Playwright to automatically explore your software, identify gaps in documentation, and draft the missing content.

Prompt: “Explore our application at [localhost URL] using Playwright. Compare what you find against the documentation in /docs. Identify any features, flows, or edge cases that are undocumented or poorly documented. Draft the missing documentation and save as new markdown files in /docs/drafts.”

Sprint Planning from Customer Signals

Connect the dots between what customers are saying and what your team should build next.

Prompt: “Read the feature request data in /data/requests.csv and the customer call summaries in /calls/summaries. Cross-reference against our current roadmap in /docs/roadmap.md. Recommend the top 5 features we should prioritize next sprint based on: request frequency, revenue impact (size of requesting accounts), alignment with roadmap themes, and estimated complexity. Output as a Linear-ready markdown spec.”

User Behavior Audit from Analytics Exports

Export your product analytics (Mixpanel, Amplitude, PostHog) as CSV. Let Claude Code find the behavioral patterns.

Prompt: “Analyze the event data in user_events_jan.csv. Identify: (1) the most common user paths from signup to first value moment, (2) where users drop off, (3) which features correlate with long-term retention, and (4) power user behaviors vs. churned user behaviors. Create a report with specific product recommendations.”

6. The Wild Ones

Use cases that don’t fit neatly into a function but are too good not to share.

Analyze Your Own Meeting Behavior

Dan Shipper’s personal favorite: download all your meeting recordings, and ask Claude Code to tell you where you’re avoiding conflict.

Prompt: “Read all meeting transcripts in /meetings. Identify every instance where I subtly avoided conflict, deferred a hard conversation, or softened critical feedback. For each, describe what happened and suggest what I could have said instead. Be honest.”

Build a Physical Object

John Conneely used Claude Code as a DIY subagent to design a slide tower for his son, complete with measurements, materials list, and step-by-step instructions.

Prompt: “I want to build a [describe physical project. e.g., standing desk, bookshelf, kids’ play structure]. I have access to standard lumber and basic power tools. Create: a materials list with quantities and dimensions, step-by-step build instructions, a cut list, and safety considerations. Include a simple diagram if possible.”

The finished product:

Expense Report from Receipt Photos

Dump 40 receipt screenshots into a folder. Walk away. Come back to a categorized expense report.

Prompt: “I have receipt screenshots in this folder. Create an Excel spreadsheet with columns: Date, Vendor, Category (meals, travel, supplies, other), Amount, and Description. Extract all information from the receipt images. Total by category at the bottom.”

Create a Self-Improvement Feedback Loop

Gang Rui created a slash command that analyzes journal entries alongside Git commits to spot gaps between intentions and actions, like a personal COO.

Prompt: “Analyze my journal entries in /journal (last 7 days) and my recent Git commits. Spot gaps between what I said I’d do and what I actually did. Suggest 3 system improvements to help me follow through better next week. Be specific and actionable.”

Voice Notes to Published Content

Voice-record ideas during morning walks, feed the raw audio notes into Claude Code, and get back organized research themes, a draft article in her exact voice, and LinkedIn posts ready to publish.

Prompt: “Read the voice note transcripts in /voice-notes. Organize the ideas into coherent themes. Then write a 1,200-word article based on the most compelling theme, match the writing style in /style-guide/my-samples. Also create 2 LinkedIn posts from the article. Save everything to /drafts.”

A Framework for Getting Started

If you’re feeling overwhelmed by the possibilities, here’s a simple decision tree:

Start with a task you already hate doing manually. Tax invoices, file organization, and data cleanup.
Create a dedicated folder. Don’t give Claude Code access to your whole machine. Scope it to a workspace.
Describe what you want in plain English. No special syntax. Just tell it what to do as if you were delegating to a smart intern.
Review the plan before it executes. Claude Code will tell you what it’s about to do. Read it, approve it, then let it run.
Graduate to GTM workflows. Once you’re comfortable, move to the sales, marketing, CS, and product use cases above.

Tag @GTMnow so we can see your takeaways and help amplify them.

OpenAI raises $110B in one of the largest private funding rounds in history. The new funding consists of a $50 billion investment from Amazon as well as $30 billion each from Nvidia and SoftBank, against a $730 billion pre-money valuation.

Profound raised $96M at a $1B valuation to help brands understand and influence how AI models like ChatGPT and Gemini talk about them. As discovery shifts from links to answers, they’re positioning themselves as the system of record for AI visibility. If you care about distribution, this is one to watch.

Anthropic introduced Claude Code Security in limited research preview. It scans entire codebases for vulnerabilities and suggests targeted patches for human review, designed to surface issues traditional security tools often miss. As AI writes more code, guardrails like this move from nice-to-have to required.

GTM 180: BREAKING: Inside Nooks’ Launch: Why AI-Native Sales Tools Are Challenging Legacy Platforms, with Co-Founder & CEO Dan Lee

Get a sneak preview here. For the full thing, listen on Apple, Spotify, YouTube or wherever you get your podcasts by searching “The GTMnow Podcast.”

Manager, Customer Success at Owner (Remote - US or Canada)
Mid Market Customer Success Manager at Seso (Remote - US)
Customer Success Manager, Strategic at Vanta (Hybrid - London, UK)
Enterprise Customer Success Manager at Noibu (Hybrid - Ottawa, ON)
Growth Development Representative at Northbeam (Hybrid)
1. Los Angeles, CA
2. San Francisco, CA

See more top GTM jobs on the GTMfund Job Board.

Upcoming events you won’t want to miss:

Funnel ‘26: March 5, 2026 (Austin, TX)
Spryng (for marketers): March 24–25, 2026 (Austin, TX)
MicroConf 2026: April 12–14, 2026 (Portland, OR)
SaaStock USA: April 15–16, 2026 (Austin, TX)
Forrester B2B Summit: April 26–29, 2026 (Phoenix, AZ)
SaaStr Annual: May 12–14, 2026 (San Mateo, CA)
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
INBOUND: September 16–18, 2026 (Boston, MA)
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
Customer Success Week: October 5-9, 2026 (NYC, NY)
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)

Some GTMnow community (founder, operator, investor) love to close it out - we appreciate you.

Subscribe now