Google is making Tasks in Gemini free in the US, Japan, Korea, Canada, Australia and India. The update runs on Gemini 4 Hydrogen, but the new calling, payment and messaging features have a narrower rollout: they start in the US.
In its 17 November announcement, Google describes a personal assistant that can work through websites and connected apps, rather than just explain what to do. People switch from Chat to Task on Gemini’s home screen to assign a job.
Calls, shopping and messages start in the US
Gemini can call businesses, though calling is not available in Indiana, Louisiana, Minnesota, Montana or Nebraska. Google says the assistant identifies itself as AI, businesses can opt out, and it needs permission to share payment or private account information.
A new Google Password Manager integration gives it access to sites with the user’s permission. Google Pay support lets it make purchases using the Agent Payments Protocol, or AP2. Google lists Mastercard and Visa support, with American Express coming later. Users can set a hard spending cap, and receipts and tracking information go to Google Wallet. Integration with the Universal Commerce Protocol is still planned, not part of this launch.
The RCS texting option requires a US phone number on an eligible mobile carrier. Setup is in Gemini’s Channels settings on Android and iOS. Email access also starts in the US, through [email protected].
Those new channels do not mean every job ends without a handoff. In the launch post, Google executive Josh Woodward describes texting Gemini to add groceries to an Instacart order, then completing the order in Instacart himself. His examples of booked appointments and calendar updates are Google’s demonstrations, not independent tests.
Permission before purchases
Google says Gemini is designed to obtain confirmation or pre-approval before clicking Buy, and to ask questions when instructions are unclear. Routine actions can be set to Always allow.
According to Google, tasks run in a sandboxed, hardware-isolated virtual machine dedicated to the user’s account. It also describes prompt-injection filters and Safe Browsing checks. These are the company’s stated safeguards, not a guarantee that a task will be completed correctly.
Google acknowledges that tasks can fail or miss the intended result. Some websites still require users to take over for the final click. The announcement gives no date for completing the six-country rollout or bringing the new US-first features elsewhere.
Banner: original editorial composition using Google’s official Gemini icon. Google and Gemini are trademarks of Google LLC. This publication is not affiliated with or endorsed by Google.
Discover more from Sulat AI & Tech News
Subscribe to get the latest posts sent to your email.




