Google said it is expanding access to the Task mode in Gemini. The mode lets the AI agent carry out multi-step jobs on the user's behalf.

What it can do

  • Coordinate phone calls
  • Log into websites
  • Make purchases on the user's behalf

The move shows Gemini shifting from a chat tool that produces answers to an agent that acts inside the browser. Other major model makers are heading the same way, handing agents tasks such as filling forms, handling accounts and completing payment steps.

Access and limits

Google did not detail the scope or timing of the expansion. Because purchases and logins require handing payment details and credentials to an agent, security and approval mechanisms become critical.

The step is part of Google's push to make Gemini more visible in everyday tasks. Which markets get the feature, and under what conditions, remains unclear.