Chrome as AI Co‑worker

Google is turning Chrome from a web browser into a workplace agent that can read what is open on your tabs and carry out tasks for employees. (techcrunch.com) Google announced the changes on April 22 at Cloud Next 2026, where it framed the broader push as building an “agentic enterprise” around Gemini. The company said nearly 75% of Google Cloud customers now use its artificial intelligence products. (blog.google) The core feature is called Auto Browse. Google said it lets Gemini use the live context from open tabs to handle multi-step work such as booking travel, entering data, and scheduling meetings, with a person still required to review and confirm the final action. (techcrunch.com) In plain terms, Chrome is being used like a remote control for web apps employees already use all day. Google said Gemini can pull details from a Google Doc, a company website, and a customer relationship management system to set up a sales lead inside the browser. (zdnet.com) Google is also adding a persistent Gemini side panel so the assistant stays available while a worker keeps the main tab open. In the consumer version, Google said the panel can compare information across tabs, summarize product reviews, and help coordinate calendars without switching windows. (blog.google) The company tied those workplace tools to other Google services, including Gmail, Calendar, Maps, YouTube, Shopping, and Flights. Google said Chrome can use those connected apps to pull event details from an email, suggest flights, and draft a follow-up message with arrival times. (blog.google) Another new feature, called Skills, lets workers save repeated prompts or workflows and run them again from the Gemini panel. Google said Workspace users can trigger those saved routines with a slash command or the plus button, and most still require confirmation before they run. (techcrunch.com) Google is also pushing image editing into the browser window itself. It said Gemini in Chrome can use its Nano Banana image model to modify images on the page without downloading the file or opening another app. (blog.google) The security pitch is aimed at the same trend Google calls “Shadow AI,” where employees use outside artificial intelligence tools without company approval. Google said Chrome Enterprise Premium is expanding its monitoring so information technology teams can spot unsanctioned tools, compromised extensions, and what it calls anomalous agent activity. (zdnet.com) Google said the first rollout is limited: Auto Browse in Chrome Enterprise is initially available to Google Workspace users in the United States and can be turned on through policy controls. The company also said Workspace prompts are not used to train its artificial intelligence models. (techcrunch.com) The shift leaves Chrome doing more of the work that once happened in separate software tabs, browser extensions, and chatbots. Google’s message at Cloud Next was that the browser itself is becoming the place where employees ask, approve, and track artificial intelligence actions. (blog.google)