Listen on YouTube // Apple // Spotify

Thank you to our sponsor partner:

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 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 you’ll learn from

Exa just passed $100M ARR, growing nearly 10x per year for almost three years. Marcus Holm, Chief Revenue Officer there after serving as President at LaunchDarkly, breaks down how a company building “what Google is for humans, but for AI agents” scaled that fast.

Marcus explains why he describes Exa as Google search for AI agents, why co-founder Will’s prediction (that 2026 would be the year agents search more than humans) came true, and what that shift means for SEO, ads, and the role of the salesperson. He gets tactical on selling to developers who don’t want to be sold to, building a consumption-based comp plan where “the sales cycle never stops,” why executive level alignment and multithreading decide most deals, and how he hires and ramps reps in a market where go-to-market talent is harder to find than engineers.

What you’ll learn

  • Why Marcus joined Exa: the criteria he uses (market size, likelihood of winning it, culture, voice of the customer)

  • Why over half of web traffic is now agents and bots, and what that does to SEO and ads

  • What Exa is in one line: “what Google is for humans, Exa is for AI agents”

  • How the salesperson’s role shifts in a bottoms-up, consumption world (from selling to being a “jungle guide”)

  • Who the real customer is when agents search: the human, the agent, and the company paying the bill

  • How to structure comp in a consumption model, where “you’re always on trial”

  • The secret to selling to developers: curiosity, credibility, and earning the right before pitching

  • The three attributes Exa hires for (humility, urgency, agency), and why he watches the first 1-2 weeks

  • His favorite interview question: “share a time you overcame the odds”

  • Why executive alignment and multithreading are the biggest unlocks in enterprise deals

  • Why go-to-market now out-hires engineering at the top AI companies, and the AI tools his team actually uses (Devin, role-play avatars)

Episode highlights

1:12 – Welcome, and why Marcus joined Exa after LaunchDarkly

2:50 – His criteria for assessing companies: market, winner, culture

3:36 – Meeting the founders, and the “never-never land” office energy

6:16 – Passing $100M ARR at ~10x per year, and how they celebrated

7:42 – Will’s prediction: 2026, the year agents search more than humans

8:43 – What Exa is: “Google for AI agents”

9:17 – What agent search means for SEO and ads

10:20 – How the salesperson’s role changes in a bottoms-up world

11:43 – Five years out: will agents buy software on their own?

13:00 – Who the end customer really is: human, agent, and company

15:29 – Structuring comp in a consumption model

17:01 – Why consumption is the clearest form of “always be selling”

18:36 – Scaling go-to-market: the PayGo-to-enterprise motion

20:27 – The secret to selling to developers

24:42 – How to find the right salespeople (humility, urgency, agency)

27:00 – Watching the first two weeks, and how reps ramp

28:10 – How late is too late to let a rep go

29:40 – The interview question that predicts performance

31:53 – What the go-to-market looks like now, post-$100M

34:35 – Lessons from LaunchDarkly and Cloudera: tech isn’t enough

36:05 – The biggest unlock: executive level alignment

38:20 – Teaching multithreading, and why it correlates with tenure

40:38 – AI usage on the team: Devin, role-play avatars, self-serve analytics

44:05 – Why go-to-market now out-hires engineering

47:45 – The biggest surprise at Exa: decisions happen on Slack

49:05 – How agent search queries keep changing, month to month

53:21 – The most underrated GTM advice: practice what you preach

54:15 – The best advice he got: “remember who you are”

55:52 – Where to find Marcus and Exa

Key takeaways

1. Agents searching the web has officially overtaken humans, and it breaks the old monetization model.

Co-founder Will predicted 2026 would be the year agents search more than humans, possibly by 1,000 to 1 eventually, and Marcus says he nailed it: per Cloudflare, over half of web traffic is now agents and bots, and agent query volume has officially passed human volume. Exa’s one-line pitch is “what Google is for humans, Exa is for AI agents”, agents can’t use consumer Google, they need an API or MCP to run searches. The knock-on effect: agents want a clean experience (no SEO, no ads), so the ad-based search model faces an innovator’s dilemma, and Exa is betting on consumption pricing instead. Human query volume will keep shrinking as assistants do the searching for us.

2. In a bottoms-up, consumption world, the salesperson becomes a “jungle guide,” and the sales cycle never ends.

With low-friction, self-serve adoption (developers can start with no API key), the land is mostly product-led; the sales role shifts to expansion, reading adoption patterns and telemetry to tell customers “I noticed you’re doing X, have you tried Y,” and connecting them to peer use cases. Marcus frames consumption as the purest form of “always be selling”: there’s no big upfront perpetual license and sometimes no renewal event, so “you’re always on trial” and must keep earning it through SLAs, shipping, and pricing. Exa’s comp plan reflects this, rewarding both the land and the expansion, with higher rates on contractually committed revenue than pay-as-you-go, to pull customers toward predictable commitments (like an AWS EDP with volume discounts).

3. Selling to developers is a “human hacking exercise”: earn the right before you pitch.

Developers are an anti-sales persona, they don’t want the “let’s grab coffee” cold ask. Marcus, who’s spent most of his career on this persona (without being a CS grad), wins by leading with value: pointing them to docs, flagging where similar teams get stuck, introducing peers on the same journey, playing matchmaker and thought leader (classic Challenger). The unlock is credibility, you have to build enough technical acumen to use the right acronyms and architecture terms so you don’t sound like “my name’s Marcus, I’ll take it from here.” Once you introduce a genuinely new perspective and ask good questions, the guard drops and they’ll finally engage on discovery and qualification.

4. Hire for humility, urgency, and agency, and the first two weeks tell you almost everything.

Exa (per CEO Will) screens for those three attributes, plus technical aptitude, grit, and a track record near the top of the team. Marcus’s favorite interview question: “share a time you overcame the odds”, vanilla “I almost lost a deal then won it” answers signal transactional thinking, while the underdog stories (homelessness, illness, working at 14) are the people he “gobbles up,” because evangelical selling in a build-phase startup demands that grit. Post-hire, early behavior is the tell: do they blaze through onboarding and binge Gong calls, or cut corners? “How you do anything is how you do everything.” On letting people go, “clarity is kind”, set explicit onboarding gates, assess will first (you can’t teach motivation), and invest in coaching skill gaps only for high-will people.

5. The product-vs-GTM bottleneck has flipped, and relationships are the durable moat. Marcus’s career-long lesson across LaunchDarkly and Cloudera: technology is never enough; “build it and they will come” always hits a wall, and you have to pair great tech with a value-selling execution model. Now coding agents make one engineer as productive as 5-10, so engineering can scale slowly while go-to-market demand is “endless, limited supply”, Exa went from almost all-engineering at 60 people to GTM now bigger than engineering at ~200, and GTMnow’s own data showed GTM was by far the largest hiring category at both OpenAI and Anthropic. His biggest deal-level unlock is executive level alignment and multithreading: most lost deals trace back to being “single-threaded down low,” so get higher, build the business case, and use AI org-charting tools to map stakeholders. Because feature parity is now easy to reach, execution and relationships become the real differentiators, the “knife fight” that still needs humans.

Marcus Holm

Sophie Buonassisi (Host)

GTMnow

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.

The media brand, GTMnow, shares how the best companies get built, told from every perspective: founders, operators and investors, at every stage of company building. GTMnow is owned and operated by VC firm, GTMfund.