Google puts Gemini into Chrome’s sidebar and adds agent-style auto-browse for subscribers

Google is tightening its grip on the browser-as-assistant trend by baking AI deeper into Chrome, shifting Gemini from a pop-up to a persistent sidebar and previewing agent-style automation that can traverse sites and complete tasks.

Gemini in Chrome moves to a persistent sidebar

The most immediate change is usability: Gemini now lives in Chrome’s side panel, so teams can query the current page or jump across open tabs without losing their workflow. Google also demoed a multi-tab behavior where tabs opened from the same source page are treated as a “context group,” which can help when comparing options across multiple product pages or pricing tabs.

This matters for marketers and e-commerce operators who routinely work with messy tab stacks (competitor research, SKU comparisons, offer validation). Instead of copying snippets into a separate chat window, the assistant is designed to work with what’s already in the browser.

The sidebar rollout expands beyond Windows and macOS to Chromebook Plus devices. Chrome remains the largest browser by market share, per global browser usage data from StatCounter, so distribution is the strategic advantage here.

Agentic auto-browse and image remixing target automation and commerce workflows

Google is also adding a “Nano Banana” integration that can modify an image using another image or product found while browsing—positioning it as lightweight creative iteration inside the browser.

More consequential is “auto-browse,” an agentic feature (software that can take actions, not just answer questions) aimed at handling end-to-end tasks like purchasing an item and hunting for discount codes. Google says the agent will pause for user approval on sensitive steps like logging in or finalizing a purchase, and the company previously indicated these flows can use Chrome’s password manager and saved payment methods without exposing those details to its models.

In early testing, Google says users tried it for appointment scheduling, form-filling, collecting tax documents, sourcing service quotes, and filing expense reports. Auto-browse is launching first for US subscribers on AI Pro and Ultra plans; the “personal intelligence” connection to Gmail, Search, YouTube, and Google Photos is expected in coming months.

For B2B teams, the takeaway is clear: browser-native agents could compress research, procurement, and admin work—if reliability improves beyond controlled demos.