Relay shut down this week. Automation startups fold constantly, so that part isn’t news. What deserves attention is where its founder landed: not another startup, not a stealth AI lab, but Google’s Chrome team.
That single move says more about the AI automation market in 2026 than the shutdown itself.
The announcement, stripped down
Jacob Bank, Relay’s founder and CEO, confirmed the closure this week. Paying customers lose access on September 14. Free users already lost theirs on August 15. Relay signaled the wind-down back in July, so this week mostly closes the loop.
Bank isn’t taking a break. He’s rejoining Google as VP of Product for Chrome, running product and developer relations for the browser. Some of his Relay team is following him in.
Bank knows this building well. Google acquired his previous startup, a scheduling app called Timeful, back in 2015. He then ran product for Gmail, Google Calendar, and Google Chat before leaving to start Relay in 2021. The pitch was simple: become the next Zapier, built AI-native from day one.
That pitch didn’t survive contact with the market.
Why this reads as a signal, not an obituary
Founders don’t usually get offered a VP seat running product for one of the world’s most-used pieces of software unless the acquirer believes the underlying thesis was sound. Just poorly distributed.
That’s the real story. Relay’s problem probably wasn’t the AI. It was the surface. A standalone workflow tool has to convince businesses to adopt a new app, migrate processes into it, and trust it with sensitive tasks like document drafting and copyediting. A browser skips that step entirely. Chrome already sits inside almost every knowledge worker’s day, so any agent living there starts with distribution most startups spend years and millions trying to buy.
Bank made the point himself, calling Chrome “a perfect place to collaborate with agents.” A founder saying that in public is close to admitting the product was right and the venue was wrong.
He’s not entering an empty field, either. Perplexity already forced the issue by taking its own agentic browser free, betting that in-browser AI assistance beats a bolt-on extension. Chrome has spent the past year answering that bet with tighter Gemini integration of its own, and the competitive pressure between the two is part of why Google wanted someone who already knows how businesses actually adopt automation, not just how engineers build it. We broke down how Perplexity’s free Comet browser raised the stakes on that exact question earlier this year.
Google is buying conviction, not just headcount
This isn’t an isolated hire. Gemini already sits inside Chrome as an agentic browsing assistant, and Google recently reported that Gemini crossed 1 billion users. Search itself has been restructured around Gemini, a shift that’s reshaping how publishers and the rest of the web compete for visibility.
Bringing in a founder who already tried and failed to solve workflow automation as a standalone company is a cheap way to import hard-won lessons. Bank spent years learning which automation workflows businesses actually want, which ones they’ll pay for, and which ones die on adoption. Google doesn’t have to run that experiment again. It just absorbed the results.
This fits a pattern that’s shown up across the industry over the past year, one tracked closely as “Acquihire 2.0”: Big Tech companies increasingly skip the messy, antitrust-exposed process of buying a company outright. They pull in the founder and the core team through a quieter deal, let the original product wind down, and absorb what the team learned. Windsurf’s team landed at Google. Scale AI’s leadership moved to Meta. Character AI’s founders returned to Google as well. Relay fits the same shape.
For the wider AI tooling market, that’s the uncomfortable part. A wave of AI automation startups launched between 2021 and 2023 on the bet that AI-native workflow tools could out-execute incumbents like Zapier. Some will still succeed independently. But Relay’s ending points to a second, quieter outcome becoming just as common: build the right idea, prove the workflows work, then get folded into the platform that already owns the browser tab.
What to watch next
Bank promised more details soon on how Chrome will let people “work with AI… to get things done.” That likely means deeper agentic task execution built directly into the browser, not a bolt-on assistant but automation as a native feature.
If it ships the way Bank’s language suggests, the real casualties won’t be the automation startups that shut down. They’ll be the ones that survive independently and spend the next two years competing against a browser that does, natively and for free, what they still charge a subscription for.
Related: Grok Bot Isn’t Your AI Assistant. It’s Your Next Junior Hire.
