Google announced on January 28 that Chrome—used by over 3 billion people worldwide—can now autonomously navigate websites, compare options, fill forms, and execute purchases through Auto Browse, a new agentic AI feature powered by Gemini 3. The update also brings a persistent side panel for instant AI assistance, on-device image generation through Nano Banana, and deep integrations with Gmail, Calendar, and other Google apps.
That shift, from passive browsing to active task execution, couldn’t come at a more critical time. Chrome still commands 70% of the browser market according to StatCounter, but that dominance is under siege. OpenAI’s Atlas browser launched in October 2025 with ChatGPT baked into every interaction, briefly triggering a 2% drop in Alphabet shares. Perplexity’s Comet browser promises “agentic search.” Microsoft Edge has Copilot. Opera added AI assistants. The competition isn’t just heating up—it’s redefining what browsers should do.
Google’s response: turn Chrome itself into an AI platform rather than fragment its massive user base with a separate browser. The stakes are enormous. Chrome isn’t just a product—it’s the gateway to Google Search, the foundation of its advertising empire, and the entry point for billions of daily web interactions.
Louis Eriakha
What Auto Browse actually does
Auto Browse handles multi-step workflows that previously required constant human input. The feature can navigate across websites, interact with forms, add items to shopping carts, compare pricing, apply discount codes, and execute transactions—all while you monitor progress in real time through Chrome’s new side panel.
Google showed exactly how this works by having the AI plan a Y2K-themed party from scratch; it analysed a reference photo, searched Etsy for matching decorations, kept everything under a $75 budget, and even applied discount codes at checkout. Beyond shopping, it acts as a hands-on assistant for tedious tasks, like comparing flights and hotels across different dates to find the cheapest weekend trip or pulling info from your PDFs to fill out complex forms automatically. It can even take on bigger projects, like managing an apartment search on Redfin with your specific filters or rounding up tax documents scattered across different accounts so you don’t have to hunt them down yourself.
The system pauses before sensitive actions. Making a purchase? Auto Browse stops and asks you to confirm. The same goes for posting to social media; it waits for your approval. You see each step it takes in the side panel, and can intervene or take over manually at any point. It’s autonomous, but not uncontrolled.
Auto Browse can access Google Password Manager with permission, meaning it can log into websites on your behalf to complete tasks. That capability raises obvious security concerns, which Google addressed by introducing what it calls “entirely new defences” for agentic browsing. The company published technical details in a December 2025 security blog outlining protections against phishing, credential theft, and prompt injection attacks.
Here’s the catch: Auto Browse is currently rolling out in preview exclusively to Google AI Pro ($19.99/month) and Google AI Ultra ($249.99/month) subscribers in the U.S. Free Chrome users don’t have access yet, and Google hasn’t specified when the broader rollout will happen.
How the rest of Gemini in Chrome works
The side panel is the most visible change. Click the Gemini icon in Chrome’s top-right corner, and a persistent AI assistant opens on the right side of your screen. Unlike the previous floating window, it stays anchored across tabs and can either focus on your current tab or pull context from multiple tabs simultaneously.
It’s free for all U.S. Chrome users with language set to English, rolling out now to MacOS, Windows, and Chromebook Plus devices. No subscription required. Early adopters report using it to compare products across e-commerce sites, summarise customer reviews from different sources, and consolidate travel itineraries without tab-switching. Google Labs VP Josh Woodward said he uses it daily for faster decision-making across multiple tabs.
Nano Banana brings on-device image generation directly into Chrome. See a photo of a living room and want to visualise different furniture? Type a prompt in the side panel. Need to turn spreadsheet data into an infographic? Same process. The feature transforms images on any webpage without requiring downloads or uploads, leveraging Gemini 3’s multimodal capabilities to understand both text and visual context. Available to all Gemini in Chrome users starting January 28.
Connected Apps integration links Gemini with Gmail, Calendar, YouTube, Maps, Google Shopping, and Google Flights. Ask Gemini to find conference details in an old email, cross-reference flight options, and draft a message confirming arrival times—all within one conversation. These integrations are opt-in through Gemini Settings. Chrome VP Parisa Tabriz wrote in the announcement, “You can keep your primary work open on one tab while using the side panel to handle a different task.”
Personal Intelligence, which launched in the Gemini app earlier this month, will arrive in Chrome in the coming months. This allows Chrome to remember context from past conversations and apply preferences automatically. Set dietary restrictions once, and Gemini applies that filter when suggesting recipes. Establish budget limits, and it factors those into shopping recommendations. All opt-in. All deletable.
Mobile support is coming. Android users can access Gemini in Chrome by long-pressing the power button. iOS users will get tab-aware queries built into the Chrome app soon, though Google hasn’t specified exact timing beyond “coming soon.”
Why Google is betting everything on Chrome
The competitive pressure is intense and accelerating. OpenAI didn’t just launch a browser—it launched a fundamentally different way of interacting with the web. ChatGPT in Atlas isn’t a sidebar feature; it’s the core interface. Users describe queries, and the browser executes them. That’s not evolution. That’s replacement.
Google has two options: build a competing AI-native browser and risk fragmenting Chrome’s 3 billion users, or transform Chrome itself into an AI platform. It chose the latter. Auto Browse, Gemini integration, and Connected Apps are the opening moves in that transformation.
The company is also preparing the infrastructure. Chrome will support Google’s Universal Commerce Protocol (UCP), an open standard co-developed with Shopify, Etsy, Wayfair, and Target. This protocol standardises how AI agents interact with e-commerce platforms, enabling Auto Browse to execute transactions across different sites without custom integrations for asksone.
That’s critical because agentic commerce—AI that completes purchases on your behalf—only works if platforms cooperate. UCP provides the framework. Whether competitors adopt it or build their own standards will shape how the agentic web develops.
The numbers are already shifting. Chrome’s market share held at 70% through 2025, but that stability masks underlying pressure. Atlas adoption is accelerating. Edge with Copilot is gaining ground in the enterprise. Perplexity Comet is attracting power users frustrated with traditional search. Google’s window to respond is narrowing.
Whether Chrome can defend its dominance depends on execution. With Auto browse working across thousands of different websites, it needs to handle different layouts and security checks without getting confused. You shouldn’t have to constantly step in to fix things. It has to feel safe, not risky. And it needs to reach everyone—not just paid users—before competitors corner the market.
For now, Chrome is no longer just a browser. It’s becoming an operating system for the web—one that browses, shops, books, and works on your behalf. Whether that’s enough to hold off the competition is the question shaping the next era of the internet.
Ogbonda Chivumnovu
