Replit CEO: AI agents reshape work, don't replace it

Amjad Masad, CEO and co-founder of Replit, used a Thursday appearance at TechCrunch Disrupt in San Francisco to describe what he calls the “self-driving company” — a model where AI agents handle much of the day-to-day grunt work inside a business while people focus on judgment and taste.

Masad said Replit’s own operations changed after its AI agents became more reliable sometime earlier this year. “The company is sort of self-improving, so a lot of the work that’s going on is driven by agents,” he said. When a support ticket arrives at Replit, an agent first triages it by reviewing GitHub issues, the company’s codebase, and Linear tickets, so that by the time a person sees the ticket, it already carries the context needed to decide how to proceed. If that person decides to ship a fix, deploying another agent to write and apply the patch is a single click.

Masad said human judgment still sits at the center of that loop, and he doesn’t expect that to change soon. He said the shift isn’t eliminating jobs so much as changing what people spend their time on: at Replit, that has meant more whiteboarding and more discussion of architecture and how to build things, as staff spend more time deciding what to work on rather than doing the manual work itself.

He added that “all of our jobs” changed over the past year, and that a large part of the public still doesn’t realize that change is coming, which he expects will surprise many. He said conversations Replit has with other businesses increasingly focus on how AI can make employees more productive and, he argued, happier — because agents can remove boring work and leave room for more creative tasks. Masad said he believes Silicon Valley has not done enough to articulate that more positive framing of AI’s effect on work, and that Replit hopes to help make that case.

Why it matters

Masad’s account suggests that as AI agents absorb routine operational work, companies may need to invest more in human judgment and strategic thinking rather than simply cutting headcount — a framing that could reshape how founders talk about AI’s role at work.