Google adds Gemini sidebar and auto-browse agent to Chrome for shopping and task automation

Google is moving faster to keep Chrome competitive with a new wave of AI-first browsers, adding a persistent assistant sidebar and an “auto-browse” agent designed to complete web tasks for you.

Gemini in Chrome shifts from pop-up to persistent sidebar

The biggest UX change is that Gemini now lives in a sidebar instead of a floating window. That matters for day-to-day workflows: you can keep the assistant visible while reading a page, scanning a product listing, or jumping between tabs. Google says the sidebar can answer questions about the current site and also reason over other open tabs.

A useful detail for buyers: when you open multiple tabs from one starting page, Chrome treats them as a single “context group.” In practice, that makes price and feature comparisons less manual because the assistant can reference related tabs without you re-explaining what you’re comparing.

This rollout expands availability beyond Windows and macOS to Chromebook Plus devices. Google is also leaning on Chrome’s scale; StatCounter lists Chrome as the world’s largest browser by market share (StatCounter).

Auto-browse and personal data connect Gemini to real workflows

For B2B marketers and e-commerce operators, the more consequential update is automation. Google is introducing auto-browse, an agentic mode (an assistant that can navigate websites and take actions) that can traverse sites on your behalf—example use cases include finding an item, applying a coupon, and proceeding through checkout. Google says the agent will pause for sensitive steps like logging in or confirming a purchase.

In the coming months, Chrome will also support Gemini’s personal intelligence capability, which connects to your Gmail, Search, YouTube, and Google Photos so you can query your own data or draft/send an email without switching apps. Separately, Chrome is adding a “Nano Banana” image feature to modify an existing image using another image or product you find while browsing.

Auto-browse is rolling out first in the US to AI Pro and Ultra subscribers. Browser agents remain error-prone in real-world browsing, but early testers reportedly used the feature for forms, appointment scheduling, collecting tax documents, quotes, and expense reports—workflows that map directly to ops and growth teams.

Google’s bet is clear: keep Chrome as the control plane for AI workflows, not just a window to the web.