SAN FRANCISCO, January 30, 2026, 02:05 PST
- Google is rolling out “auto browse” in Chrome, letting Gemini take actions across websites on a user’s behalf.
- The feature is launching as a U.S. preview for paid Google AI Pro and AI Ultra subscribers.
- The update also deepens Gemini’s place in Chrome with a pinned side panel and new tools such as Nano Banana image editing.
Google is rolling out an “auto browse” feature in Chrome that lets its Gemini 3 AI take over a tab and complete multi-step online tasks, from researching travel options to filling out forms. The tool is starting as a U.S. preview for paying Google AI Pro and AI Ultra subscribers, and Google said Chrome will support its Universal Commerce Protocol (UCP), an open standard for agent-led shopping co-developed with Shopify, Etsy, Wayfair and Target. 1
The move puts Google’s browser into the fast-moving market for “agentic” AI — systems that can click and type on a user’s behalf, not just answer prompts. A wave of AI-first browsers from companies including OpenAI and Perplexity has been pitched as a way to replace Chrome-style tab juggling, and TechCrunch called Chrome the world’s largest browser by market share. 2
Chrome’s AI push also lands as U.S. regulators keep pressure on Google over search. An antitrust judge recently rejected the Justice Department’s request that Google be forced to sell Chrome, pointing to AI-driven competition from smaller rivals including OpenAI and Perplexity, AP reported. 3
In Chrome, auto browse runs through a Gemini side panel that now stays anchored to the right of the screen and can draw on apps such as Gmail, Calendar, Maps, Google Shopping and Google Flights. “Gemini can dig up that old email with event details,” Google wrote, as one example of how the integrations could work, while a new tool called Nano Banana lets users edit images in the browser with a text prompt; Google also plans to bring its opt-in “Personal Intelligence” memory feature to Chrome in the “coming months,” The Verge reported. 4
Bloomberg reported that the assistant can open websites and click around on a user’s behalf without leaving Chrome. Charmaine D’Silva, a director of product for Chrome, said it could “open different airline and hotel websites to compare prices” when planning a family trip. 5
In a prelaunch demo, D’Silva framed the tool as a way to offload repeat shopping: “I can now delegate to Auto Browse within Gemini to be able to go ahead and buy jackets for me.” WIRED said the system shows its steps and asks before more sensitive actions like paying or posting, while warning that agents can still be vulnerable to “prompt injection” — hidden instructions on malicious pages that try to trick the bot into doing something the user did not ask for. 6
But analysts said browser-only automation can break on complex sites and may be best suited for low-risk chores, at least at first. Pareekh Jain, a principal analyst at Pareekh Consulting, called it useful for “complex multi-step web workflows like form filling and navigation,” but warned it can be brittle on pages where the DOM — the web page’s underlying structure that the browser uses to find buttons and fields — changes often; Abhisekh Satapathy, a principal analyst at Avasant, said “in regulated environments, this can complicate audits and compliance reviews.” 7
For Google, the bet is straightforward: keep Chrome central as people shift from searching and clicking to delegating tasks to software that acts for them. The harder part is proving it can work on the open web, reliably, without users feeling they’ve handed the steering wheel to something they can’t quite control.
