From Perfect Search to the AI Stack: 10 Lessons from Exa AI CEO Will Bryk
Artificial intelligence may be evolving at breakneck speed, but many of the companies defining this era weren't built by chasing the latest trend. During a recent Enrich session, Exa AI CEO & Co-founder Will Bryk shared how his team spent years solving a seemingly niche problem—finding better information on the internet—before AI agents made that technology indispensable. Along the way, he offered thoughtful perspectives on building products, leading engineering teams, hiring exceptional talent, and navigating hypergrowth without losing focus.
Whether you're building AI products, leading engineering organizations, or thinking about the future of software, these were some of the biggest takeaways from the conversation.
Key Takeaways
1. Build around a durable problem—not today's trend.
Exa wasn't founded to power AI agents. Will started the company because he believed finding high-quality information on the internet was fundamentally broken. By solving a long-term problem rather than chasing a market trend, Exa was perfectly positioned when AI products suddenly needed a better search layer.
2. Don't try to boil the ocean—solve one valuable problem first.
Rather than attempting to index the entire web from day one, Exa initially focused on smaller, high-value datasets like companies and people. This allowed the team to create meaningful customer value long before they had the resources to crawl the full internet.
3. Small, high-agency teams outperform large teams in an AI-native world.
Will argued that AI coding agents fundamentally change how engineering organizations should be structured. At Exa, project teams of roughly three engineers create enough collaboration to improve decisions while remaining small enough to move quickly without unnecessary coordination overhead.
4. Founder-led product alignment can outperform rigid processes in hypergrowth.
Rather than relying on heavy product management processes, Will and his co-founder spend much of their time talking with engineers, reviewing work in progress, and ensuring everyone stays aligned around a shared vision. As the company scales, more formal structure is emerging—but only where it's needed.
5. Hire for passion, judgment, and culture—not just technical ability.
Will believes AI gives every employee "an army of 100 interns," making ambition and judgment more important than ever. Exa looks for people who are deeply curious, excited to build, make good decisions, and are enjoyable teammates—not simply the strongest technical candidates.
6. Culture doesn't scale automatically—it requires deliberate investment.
As Exa rapidly doubled headcount, Will found that maintaining culture required intentional effort. Regular team lunches, picnics, and founder accessibility became essential for integrating new employees and preserving the energy that defined the company in its earliest days.
7. Customer requests should shape your product—but not dictate it.
Exa avoids building one-off features for individual customers unless they support the company's broader product vision. Instead, the team looks for requests that are broadly applicable and improve the core product for many customers rather than customizing for each account.
8. Great go-to-market starts with building something developers genuinely love.
Exa's early growth came primarily through developer adoption rather than enterprise sales. As developers began building AI applications with Exa, the company identified high-intent users and expanded those relationships into larger commercial opportunities before investing heavily in outbound sales.
9. Data—not model intelligence—will become the biggest competitive advantage in AI.
One of Will's strongest predictions was that frontier models are becoming "smart enough" for most knowledge work. The differentiator will increasingly be the quality, freshness, and accessibility of the data those models can access, making information infrastructure more valuable than ever.
10. Persistence creates opportunities that are impossible to predict upfront.
Exa spent two years building "perfect search" before realizing AI agents would become their ideal customer. Will reflected that if they had focused only on short-term monetization, they likely would have quit before the market caught up. Solving an enduring problem—and sticking with it long enough—ultimately positioned the company for explosive growth.
Final Thoughts
One of the most compelling themes from Will's conversation was that many of the principles behind building enduring companies haven't changed—even as AI transforms how we build software. Focus on meaningful problems, hire exceptional people, move quickly with small teams, and stay relentlessly close to your customers. AI may be accelerating product development, but durable strategy, sound judgment, and long-term conviction remain the qualities that separate lasting companies from fleeting ones. As the AI ecosystem continues to evolve, the winners may not be those chasing the next breakthrough, but those building the infrastructure and culture that can adapt to whatever comes next.
For the complete summary of takeaways from this event, visit our membership events page on Notion.
Not a member? Apply to join and get access to all of our past event recordings.
