Key Takeaways

  • Mc Lader left his role as president of Uniswap to run Cognition Industries, building dedicated units to deploy the Devin AI agent into enterprise sectors.
  • Cognition is launching across 14 major industries where software is central, targeting traditional spaces like energy, healthcare, and airlines.
  • Instead of hiring traditional enterprise sales reps, Cognition is recruiting former startup founders who have already built and sold software in those specific verticals.
  • Enterprise AI adoption is stalling because companies treat advanced models like brilliant interns welcomed into the office without projects or assignments.
  • The company's go-to-market strategy focuses on partnering directly with incumbents to modernize legacy software systems and fix cybersecurity risks.

Enterprises Do Not Know What to Give Their AI Interns

Enterprise software buyers rushed to purchase generative AI access over the last eighteen months, but actual usage remains shallow. Companies bought licenses for staff without redesigning their operating workflows. Lader points out that companies ended up with enormous computational capability that nobody knows how to steer:

“So we'll start with go to market because there's a lot of just like showing up like you understand and know an industry, right? And I think that's really important, right? So we're kind of in this era of enterprise adoption where it's like the world's best university appeared next to your company and you welcomed in all the interns but you didn't like give them projects and like direction on what to do.”

Handing a developer an open chat prompt or an automated coding assistant does not modernize a twenty-year-old COBOL claims system. General developer tooling targets individual speed improvements on everyday tickets. Cognition wants to aim Devin at foundational technical debt, which requires understanding the specific operational reality of an airline dispatch system or a hospital records platform.

Cognition Hires Former Founders Instead of Account Executives

To bridge that gap, Cognition is skipping standard enterprise account executives. Lader, who previously founded a stablecoin company acquired by Coinbase before serving as president of Uniswap, is building pods led entirely by former startup founders:

“I was president Uniswap, started stable coin company, sold to Coinbase, and then as I was thinking about what was most exciting in the world, I felt like the cognition was one of the most powerful and exciting companies to be part of.”

His hiring mandate looks for builders who spent years selling into specific enterprise verticals:

“So initially I'm looking for people who like I said ship product and sold into an industry are opinionated about what AI now makes possible and are excited to come do that with really the full force of cognition.”

A standard enterprise sales rep pitches software features to procurement departments. A founder who spent five years attempting to sell EHR integrations to hospitals understands the exact regulatory walls, data silos, and technical debt inside the building. Putting a technical founder in charge turns an AI sales call into a systems architecture review.

Legacy Codebases Need Builders With Opinions

Cognition is targeting 14 core sectors where software runs the business but modernization has moved slowly. As Lader describes it:

“We're starting with 14 and it's basically the the 14. Yeah. So you can think of that as the 14 like largest industries in the global economy but also where software is pretty central.”

Airlines, energy grids, and hospital chains do not need more seat licenses for code completion. They have millions of lines of aging internal software that their current engineering teams cannot afford to touch without breaking core operations. Deploying autonomous agents like Devin into these systems requires deep industry context to identify which legacy systems to rewrite, how to test them against regulatory standards, and where automated modernization creates immediate business value.

What to Do With This

Stop pitching generic AI productivity metrics to enterprise buyers this week. Pick your single largest enterprise prospect, audit their public job boards for the legacy tech stack they are struggling to maintain (such as outdated database migrations or legacy Java services), and submit a concrete proposal showing the exact system your tool can refactor first.