Google Chrome can turn your desktop or Android device into an agentic AI machine

Google has made a very subtle yet significant agentic artificial intelligence (AI) push for consumers in India, with the new Gemini in Chrome’s auto browse feature now available on desktop as well as Android phones. Speaking of which, Google is also unlocking the broader Gemini in Chrome AI browsing assistant for all Android users in India—Gemini in Chrome had rolled out on Windows, Mac and Chromebook Plus computing devices, earlier this year.

Think of this as an AI assistant embedded within the Chrome web browser, which is by far the most popular web browser across platforms, worldwide and in India. According to the latest data by research firm Statcounter, Chrome has cornered about 89.07% of the market share in India, with Opera (3.93%), Apple Safari (2.82%) and Microsoft Edge (1.13%) following distantly.

It is important to make a distinction regarding availability. Gemini in Chrome on Android is available for all users, with the ability to summarise web pages, connect between Google apps such as Calendar or Gmail, and generate images using the Nano Banana model. However, the auto-browse feature on desktop and Android will be available to Google AI Pro and AI Ultra subscribers in India.

The key with Gemini in Chrome auto browse is that it can autonomously navigate websites, web pages and even online forms that need to be filled out. It can also be configured to compare products to search for deals and price changes before a user makes a purchase, book tickets for events, make restaurant reservations and schedule appointments. Basically, Gemini in Chrome auto browse will be able to perform complex, multi-step tasks on the web.

“These capabilities are built to be secure-by-design, and our models are trained to detect known threats, such as prompt injection. For added control, auto browse is designed to keep you in control by asking for confirmation before completing some sensitive tasks,” says Charmaine Dsilva, Senior Director, Product Management for Chrome, in a statement.

This clarity is important because giving an AI system extensive ability to interact with websites inherently introduces risks that are more evolved than when asking a chatbot a question. Prompt-injection attacks embedded in webpages can attempt to manipulate an agent’s instructions, while mistakes become more consequential if an AI agent can change orders or initiate transactions without confirmation.

As for how many agentic tasks you can ask Gemini in Chrome to execute, depends on the subscription tier. As things stand, Google AI Pro plan users can make up to 20 requests per day for multi-step tasks, while Google AI Ultra plan subscribers can send in up to 200 requests per day.