Google starts free Gemini Task mode rollout in six markets
Google is rolling out Gemini Task mode at no charge in six markets, with multi-step workflows for eligible personal accounts and separate limits on calling, payments and other sensitive actions.
Google has begun rolling out Gemini Task mode at no charge in the United States, Japan, South Korea, Canada, Australia and India, giving eligible users a way to delegate multi-step work across websites and connected apps. Google’s examples span business calls, appointment booking, bill analysis and online shopping, but its public documentation sets separate eligibility and confirmation boundaries for some of those actions.
The staged release is available without a Google AI subscription to adults using personal Google Accounts with Gemini History enabled; work and school accounts are not eligible. Google’s help page lists Task mode in the Gemini mobile and web apps and the Gemini app on Mac. Because the company describes the release as rolling out, availability may not have reached every eligible account in the six markets.
In Google’s launch examples, Gemini searches websites, calls a business, adds an appointment to Calendar, signs in to a utility portal to analyze a bill, compares products and puts items in a shopping cart. The accompanying Task mode documentation says the experimental feature can run as many as 15 tasks at once and is subject to compute-based usage limits.
What Gemini can complete depends on the services and permissions a user connects. Google lists Contacts, Google Pay and Wallet, Photos, Workspace apps, Search services and YouTube among the supported Google services, alongside several third-party apps. The company says completing a task may send information needed for the job, including potentially sensitive data from connected apps, to websites, apps and other third parties. Users can stop or take over browser work and clear data held by Google’s remote browser.
Google says Gemini pauses for review before sensitive actions including purchases, financial transactions, communications, data deletion, sharing sensitive information, form submission and sign-in. Users may grant permission once, for a chat where that option is offered, or more broadly, but the permissions guidance says the most sensitive actions still require confirmation each time. Sign in with Google and Google Password Manager also require approval, with saved credentials filled only after the user approves the specific site. Local browser automation requires an initial Chrome connection and confirmation for each browsing task.
For shopping, Google says Gemini can use the Agent Payments Protocol with Google Pay or Wallet and a user-set spending cap. Supported cards retain their rewards and protections, according to the company. Mastercard and Visa are participating in tokenized payments, while American Express support is described as coming soon. Google’s grocery example ends by handing the user to Instacart to complete the order, so some workflows still stop short of checkout. Universal Commerce Protocol integration is also described as a future step rather than part of the current release.
The announcement says Gemini can pay bills, but its detailed utility example shows the assistant signing in and analyzing a bill, not completing a payment. Google did not identify supported billers or describe a separate confirmation flow for bill payment in the supplied materials.
Calling has its own restrictions. Google says an automated call opens by disclosing that an AI assistant is calling for the user, and businesses can opt out of receiving such calls. The calling policy excludes emergency services, financial transactions, telemarketing and calls that share passwords or sensitive financial, health or government-ID information.
Google’s public pages leave an eligibility conflict unresolved. The launch post presents calling within the free six-market Task rollout and notes exclusions in five U.S. states. The dedicated calling page instead describes gradual U.S.-only availability requiring a Pixel 10 or Pixel 11, a U.S. SIM, a Google AI subscription, the public-beta Phone by Google app, the latest Gemini app and an English-language device. Google does not explain whether the pages refer to separate calling implementations.
A smaller group of advanced features is starting in the United States. Google explicitly identifies forwarding email to gemini@google.com and texting Gemini; texting requires a U.S. phone number on an eligible carrier. The announcement does not provide an exhaustive list separating every U.S.-first feature from the core rollout or dates for wider availability.
More news

xAI deploys five Grok Bots across accounting workflows

Perplexity launches Frontier Research, claims access to 10,000-plus GPUs

Microsoft introduces Azure Catalyst and agent infrastructure tools
