Relay Shuts Down as Google Acqui-Hires Its Team for Chrome's Agentic Ambitions

Relay, an AI-powered workflow automation startup launched in 2021 with the explicit ambition of becoming “the new Zapier,” is shutting down — and its talent is being absorbed by Google in an acqui-hire aimed squarely at the browser. On August 17, TechCrunch reported that Relay founder and CEO Jacob Bank will rejoin Google as VP of Product for Google Chrome, leading the browser’s product and developer relations teams, with other Relay staff following him over.

The mechanics of the shutdown are already in motion. Paying customers lose access to Relay on September 14, while free-tier users were cut off as early as August 15. The closure was first announced internally to users in July, making this week’s news less a surprise collapse than a planned wind-down wrapped around a talent deal.

For a startup that positioned itself as a next-generation automation platform — streamlining repetitive business tasks like document drafting, copyediting, and project management chores — the ending is bittersweet but instructive. Relay didn’t run out of runway so much as run into the reality of 2026: when the giants build your product into their platforms, independent workflow automation becomes a very hard place to live.

The founder’s full-circle journey

Jacob Bank’s career arc is the interesting part of this story. He originally joined Google in 2015 when his scheduling startup Timeful was acquired by the search giant. He then spent a little over six years inside Google, including a stint as product lead for Gmail, Google Calendar, and Google Chat — the exact suite of productivity surfaces where AI assistance has landed hardest over the past two years.

He left to build Relay, and now he’s back, one rung higher. In a post on X announcing the move, Bank framed it as a continuity of mission rather than a retreat: “I’ve spent my whole career doing one thing: building tools that help people get more done with AI, without sacrificing their personal creativity or insights. And joining the Chrome team is an ideal opportunity to bring those experiences to many, many more people.”

The more tantalizing line came next: “We have some really ambitious plans to help you work with AI in Chrome to get things done, and I’ll have more to share soon.” Combined with his description of Chrome as “a perfect place to collaborate with agents,” it reads less like a courtesy quote and more like a roadmap.

Why Chrome is the battleground

To understand why Google would buy a workflow-automation team for its browser, look at what Chrome has become in 2026. In January, Google shipped an agentic “auto browse” mode in Chrome that autonomously navigates websites to complete tasks. At I/O 2026 in May, the Chrome team unveiled a fifteen-part update package headlined by WebMCP — a framework that lets websites expose themselves as agentic toolkits — alongside a blueprint for guiding coding agents to build for the modern web.

The strategic logic is straightforward. As AI agents become primary users of the web, the browser stops being a passive document viewer and becomes an agent runtime: the place where autonomous assistants read, click, fill, and transact on a user’s behalf. Whoever controls that runtime controls the distribution layer for the entire agentic economy. Perplexity’s Comet, OpenAI’s Atlas, and a wave of AI-native browsers have been attacking from the outside; Google’s counter is to make Chrome itself the best agentic environment before the challengers gain ground.

Bank’s hire fits that plan precisely. Relay’s core competency was orchestrating multi-step AI workflows across business tools — deciding what to automate, chaining the steps, and keeping humans in the loop where judgment matters. Those are the exact hard problems of browser agents, which must act across arbitrary websites with reliability that neither pure LLMs nor brittle RPA scripts can deliver alone.

The consolidation wave behind the deal

Relay’s fate is also a data point in a broader consolidation now sweeping the AI application layer. Standalone startups that once looked like durable businesses — thin wrappers adding automation on top of foundation models — are increasingly finding their features absorbed by the model providers and platform owners. The lucky ones get acqui-hired, as Relay was. Stripe’s reported $7 billion-plus deal for AI gateway OpenRouter, reported the same week, shows the same dynamic at a different scale: distribution and platform position are the scarce assets, and everything else is being rolled up.

There’s an irony worth noting: Relay set out to disrupt Zapier, and instead both the disruptor and the incumbents now face a bigger threat — the browser and the OS becoming smart enough that dedicated automation middleware gets squeezed from both ends. Tracxn tracked Relay as a seed-stage company founded in 2021 with roughly $8 million raised, a modest sum by 2026’s standards. That capital was never going to fund a war against Google, Microsoft, and OpenAI all simultaneously shipping agent products.

What happens next

For Relay’s customers, the message is blunt: export your workflows before September 14. The company has not announced any data migration partnerships, and free users have already lost access.

For everyone else, watch Google’s next Chrome announcements closely. A VP of Product hire who talks about “ambitious plans” for AI in Chrome, arriving with a team that spent five years building agentic workflow automation, is not being brought in to maintain the status quo. Gemini has already been integrated into Chrome as an optional in-browser assistant, and Google recently reported Gemini crossing 1 billion users — the audience for whatever Bank’s team builds next is the largest of any software product on Earth.

The browser wars of the 2010s were about search defaults and rendering engines. The one now beginning will be about whose agent you trust to act on your behalf — and Google just spent an acqui-hire to make sure the answer runs on Chrome.