Google is making its Gemini Tasks feature free for users in six countries, opening up its AI-powered assistant to a much wider audience. Previously available to paid subscribers, Tasks can handle multi-step projects and recurring routines in the background, rather than simply responding to prompts. The move is part of Google’s broader push to turn Gemini from a chatbot into an assistant that can take action on a user’s behalf.
Gemini can now handle more than just prompts
Think of Gemini Tasks as handing your AI assistant a to-do list and letting it get on with things. It can build travel itineraries, schedule daily briefings using information from emails and calendars, automate repetitive research, and even manage household chores. Tasks runs in the background, sends a notification when the job is done, and lets users add instructions along the way.
Previously known as Gemini Spark, Tasks is now part of the main Gemini app and is available to free users in the US, India, Canada, Australia, Japan, and South Korea. Getting started is simple: open Gemini and switch from Chat to Task. No paid subscription is required, although some of the newer capabilities have separate availability restrictions.
Gemini is also learning to pick up the phone, quite literally. It can call multiple businesses to compare prices, check inventory, get quotes, or schedule appointments, identifying itself as an AI assistant in the process. Google Pay integration can also track product availability and price drops, then prepare purchases within a user-defined spending limit. There’s still a human checkpoint, though: users must review the itemised summary and approve the purchase, with receipts and tracking details saved to Google Wallet.

And there’s no need to keep the Gemini app open to delegate work. US users can text Gemini through Google Messages or iMessage, or forward emails and attachments to [email protected]. It can pull out relevant information, organise calendar entries, and send the completed work back privately. Between phone calls, shopping, and inbox chores, Gemini is slowly graduating from answering questions to doing the boring bits of life.
Google is joining the race to build AI agents
The update reflects a broader shift in how AI assistants are being developed. Instead of simply answering questions or generating content, companies are increasingly trying to build agents that can interact with services and complete tasks independently. Meta’s Muse, for instance, has also been positioned around taking action across connected services. However, AI agents can run into problems when websites and businesses restrict automated access. That’s a challenge Google acknowledges, too. The company says some websites may prevent Gemini from completing an action, forcing it to hand control back to the user for the final step.

Google says Gemini Tasks runs in a dedicated, hardware-isolated virtual machine for each user, with safeguards against prompt injection and unsafe websites. It also keeps users in control by requiring approval for sensitive actions, such as purchases, although Google admits Tasks can still make mistakes and some websites may not play nicely with AI agents. Making it free is a big step, but the real test is whether Gemini can reliably handle the boring bits of everyday life without needing constant supervision.
