Google rolls out Chrome AI agent that can browse and complete tasks for users

Google is introducing an artificial intelligence assistant within its Chrome browser that can open websites and navigate them on a user’s behalf, according to a Bloomberg report. The new feature, called auto browse, enables users to ask an assistant powered by Gemini to carry out tasks such as shopping or planning travel without leaving the browser, Charmaine D’Silva, a director of product at Google, said.

D’Silva explained that Chrome users will be able to plan activities such as a family trip by asking Gemini to open and compare airline and hotel websites to find suitable options. The feature is designed to reduce the need for manual navigation across multiple tabs while completing routine online tasks.

Parisa Tabriz, vice president of Chrome, said in a company blog post that early testers had already used the tool for a wide range of activities, including scheduling appointments, filling out lengthy online forms, collecting tax documents, obtaining quotes from plumbers and electricians, checking whether bills had been paid, filing expense reports, managing subscriptions and speeding up processes such as renewing driving licences, resulting in significant time savings.

The launch marks the latest effort by Google’s parent company, Alphabet Inc., to integrate artificial intelligence more deeply across its products. Earlier this month, the company said Gemini could draw on user data from services including Gmail, Search, Photos and YouTube to deliver more personalised responses. With this capability, known as Personal Intelligence, enabled on Chrome, Gemini can assist with logistical planning such as reviewing school schedules and existing appointments to determine the most suitable travel options, D’Silva said during a media call.

Chrome’s auto browse feature will initially be available to US-based AI Pro and AI Ultra subscribers and will use Google Password Manager to sign into websites on a user’s behalf. As part of the same rollout, Google is also integrating its image generation tool, Nano Banana, directly into the Chrome browser.

The company said safeguards have been put in place to ensure the agentic AI cannot take final actions, such as placing an order, without explicit user approval. Tabriz said during the call that Google is combining AI with on-device models to protect users against an increasingly complex threat landscape, including AI-generated scams and more sophisticated cyberattacks.