Summary:

Free Tasks is expanding to six countries, while Google Pay checkout, texting and email have narrower U.S. eligibility and approval requirements.

Google is bringing Tasks in Gemini to free users in six countries, but the rollout does not make every new agent feature available everywhere. The November 17 announcement covers the United States, Japan, Korea, Canada, Australia and India, with advanced capabilities starting in the U.S.

Tasks uses Gemini 4 Hydrogen to carry out work rather than only answer questions. Users switch from Chat to Task mode and describe what they want done. This is a product-access expansion, distinct from the model's launch. Google's announcement describes connected-app workflows, business calls and shopping assistance, with some websites still requiring a handoff to finish.

Which features are available where?

  • Free Tasks: rolling out in the six countries above.
  • Google Pay checkout: rolling out in English in the U.S. on web, Android and iOS for adults with personal Google accounts.
  • Texting: U.S. adults need a personal account, a verified U.S. phone number and an eligible carrier. Google's release notes cover Google Messages and iMessage.
  • Email: starting in English for U.S. adults with personal accounts. It works with Gmail, Apple Mail and Outlook, but replies privately rather than to everyone on a thread.

These extra requirements come from the November 17 Gemini release notes. A country appearing in the free Tasks rollout should not be read as confirmation that payments, texting or email are available there.

A spending cap does not replace purchase approval

Google's release notes say checkout presents an itemized summary and requires explicit one-tap and device approval before purchase. The system uses a single-use virtual card rather than exposing the actual card number to the merchant. Users should therefore distinguish setting a budget from approving the final transaction.

Business calling also has boundaries. Google's launch announcement excludes Indiana, Louisiana, Minnesota, Montana and Nebraska. Gemini identifies itself as AI, and businesses can opt out.

What changes when an assistant can act?

Google's security explanation describes hardware-isolated computing environments, checks before planning, inspection of commands and files, and a separate check on external actions. High-impact actions can still require confirmation even when an “Always allow” preference exists.

For supported sites, Google Password Manager can fill credentials without putting passwords into the model's conversation. Google lists that integration for web and Android in the U.S.; it also provides a way to clear remote logins in settings.

The practical starting point is a narrow task: name the outcome, limit the scope and state what needs approval. “Compare three delivery options and stop before ordering” is clearer than “handle dinner.” Review the destination, recipients and final price when a task reaches a confirmation screen.

Sources and image credit

Google's November 17 announcement, Gemini release notes and security documentation are linked above. The cover is an original PromptsZone rollout diagram based on those sources, not a screenshot of the Gemini interface. Google and Gemini are trademarks of Google LLC. This article is not endorsed by or affiliated with Google.