Google is expanding access to Gemini Tasks, its AI-powered agent designed to handle multi-step activities beyond ordinary chatbot conversations. The feature is rolling out free to hundreds of millions of users across the US, India, Canada, Australia, Japan and South Korea, allowing Gemini to work on tasks in the background, organise information and help users complete everyday activities with less manual effort.
The expansion represents Google’s push to turn Gemini into an AI assistant capable of taking action rather than simply answering questions. While the broader Tasks experience is expanding internationally, advanced capabilities such as calling businesses on users’ behalf and completing purchases through Google Pay are currently being rolled out with specific regional availability, particularly in the US. The development marks another step in the competition to build AI agents capable of interacting with the real world.
Google Expands Gemini Tasks To Six Countries
Google is making Tasks in Gemini available free of charge to users in six markets: the United States, India, Canada, Australia, Japan and South Korea.
The feature transforms Gemini from a conversational assistant into an agent capable of working on multi-step assignments in the background.
Instead of responding to a single prompt and stopping, Gemini Tasks can carry out a series of steps, organise information, conduct research and return results after completing the assigned work.
The rollout is powered by Google’s Gemini 4 Hydrogen model, which is designed to handle agentic workflows and tasks requiring multiple actions.
Users can switch from the regular Chat interface to Tasks within the Gemini app and assign work they want the AI to complete.
Google is positioning the feature for everyday use cases, including researching information, organising schedules, preparing travel plans, analysing documents and managing recurring activities.
Gemini Can Call Businesses And Book Appointments
One of the most notable capabilities associated with Gemini Tasks is AI-powered calling.
Gemini can call businesses to ask questions, compare services, check availability and arrange appointments on behalf of users. The technology is designed to handle conversations, wait on hold and communicate requests without requiring the user to make the call personally.
For example, users could ask Gemini to contact multiple car dealerships to compare service appointment availability, call repair businesses to obtain quotes or contact local establishments to confirm whether a service is available.
The AI can also use relevant information from a user’s calendar to help identify suitable appointment times.
Google says its AI identifies itself as an AI assistant when contacting businesses. Companies can also opt out of receiving these automated calls.
However, agentic calling remains subject to regional availability and eligibility requirements, with the current rollout focused on the US rather than being universally available in all six countries receiving Tasks.
Gemini Can Search For Products And Complete Purchases
Google is also expanding its AI shopping capabilities by allowing Gemini to help users discover products, compare options and prepare purchases.
The assistant can search for items according to a user’s requirements, including price limits, specifications and preferences.
In supported markets, Gemini can also use Google Pay to complete purchases after the user approves the transaction.
Google is introducing spending limits to give users greater control over purchases made through AI-assisted shopping. Users can specify a budget, allowing the assistant to look for suitable products without exceeding the defined limit.
The purchasing workflow is designed to include a review step before the transaction is completed. Gemini can prepare checkout details and present an itemised summary for approval rather than making purchases without the user’s consent.
Receipts and order-tracking information can also be integrated into Google’s services, including Google Wallet.
The purchasing functionality is currently subject to regional rollout restrictions, with the Google Pay agentic checkout capabilities initially focused on eligible US users.
Google Introduces More Secure AI Agent Payments
AI-powered shopping creates a new set of security challenges because assistants need to interact with payment systems, merchants and sensitive account information.
Google says its agentic payment experience uses tokenised payment credentials rather than exposing a user’s actual card details directly to merchants.
The company is working with payment networks including Visa and Mastercard, with American Express integration also planned.
The system is designed to use digital payment tokens and virtual-card mechanisms to help protect financial information during AI-assisted transactions.
Google is also introducing safeguards requiring user approval before important actions are completed, including purchases.
These controls are important as AI agents become capable of navigating websites, entering information, signing into accounts and interacting with checkout systems.
Gemini Tasks Can Sign Into Websites With User Permission
Google is expanding Gemini’s ability to interact with websites and accounts using securely stored credentials.
With user permission, Gemini can access supported login information through Google Password Manager and sign into websites when required to complete an assigned task.
This could help the AI perform activities that otherwise require users to manually log in, navigate menus, retrieve information or complete forms.
For example, an assistant could access a supported account to review information, analyse documents or perform an activity requested by the user.
However, sensitive actions remain subject to permission controls and additional confirmation steps.
The feature also introduces new security considerations because AI agents interacting with websites could encounter malicious instructions, deceptive pages or attempts to manipulate their actions.
Google says Tasks operates within an isolated computing environment and uses security protections designed to limit risks associated with AI agent activity.
Tasks Runs In An Isolated Computing Environment
Google says Gemini Tasks uses its Agent Compute Engine infrastructure to carry out work in a dedicated, hardware-isolated virtual machine.
The environment is designed to separate the agent’s activity from other users and provide additional protections when it interacts with websites and online services.
Google also uses security systems intended to detect prompt injection attempts and potentially harmful links encountered during web-based tasks.
Prompt injection is a security risk in which malicious instructions embedded in websites, documents or other content attempt to manipulate an AI agent into ignoring its original instructions or performing unwanted actions.
The risks become more significant when AI systems are able to interact with accounts, payments and other sensitive services.
Google says users can intervene while Tasks is running, provide additional instructions and approve important actions before they are completed.
The company also acknowledges that automated tasks may sometimes fail, stop midway or produce results that do not fully match a user’s expectations.
Google Wants Gemini To Become An Everyday AI Assistant
The expansion of Tasks reflects Google’s broader strategy to make Gemini more deeply integrated into everyday digital activities.
Traditional AI chatbots primarily generate answers, explain concepts, summarise information and create content. Agentic AI systems are designed to take those capabilities further by interacting with software, websites, services and digital tools.
Google wants Gemini to handle activities such as:
- Organising travel plans and researching destinations.
- Managing schedules and calendar activities.
- Comparing products and shopping options.
- Analysing emails, documents and bills.
- Researching information across multiple websites.
- Completing repetitive multi-step assignments.
- Monitoring certain tasks and returning results later.
The strategy could make Gemini more useful to people who already use Google’s ecosystem of services, including Gmail, Google Calendar, Google Wallet and Google Search.
Competition In AI Agents Is Intensifying
Google’s expansion comes as major technology companies compete to build AI assistants capable of performing tasks autonomously.
OpenAI, Anthropic, Microsoft and other companies are developing agentic systems that can use tools, interact with software and execute multi-step workflows.
The competitive focus is shifting from the quality of chatbot responses to the ability of AI systems to reliably complete useful tasks.
The companies that succeed could potentially create new ways for consumers to shop, businesses to operate and software to interact with users.
However, AI agents must also overcome challenges involving accuracy, privacy, security, reliability and user trust.
The Bigger Picture
Gemini Tasks reflects a fundamental shift in how AI assistants are being developed.
Instead of requiring users to manually perform every step after receiving an AI-generated answer, agentic systems can increasingly handle parts of the workflow themselves.
This could reshape online shopping, customer service, travel planning, administrative work and interactions between consumers and businesses.
Google’s strategy also gives it an advantage through its existing ecosystem of search, maps, payments, email, calendars and Android devices.
The challenge will be ensuring that AI agents can operate reliably while maintaining user control over sensitive information and financial transactions.
Looking Ahead
Google is expected to continue expanding Gemini Tasks into more workflows, services and potentially additional markets. The success of the feature will depend on how reliably it completes assignments, how easily users can supervise its actions and whether businesses allow AI agents to interact with their systems. Wider availability of calling and purchasing features will also depend on regional restrictions, payment infrastructure and regulatory requirements.
The broader AI industry is moving toward assistants that can take action rather than simply provide information. If Google can make these capabilities secure, dependable and easy to use, Gemini Tasks could become an important part of how consumers interact with digital services. The shift from asking AI for answers to delegating real-world tasks could represent one of the biggest changes in the next phase of consumer AI.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.



