
Tech • AI • Robotics • Game
⏱ ≈ 3 min read · 15-min video — 12 min saved
Google has launched Gemini Task, a free personal AI agent in several major markets that can handle messages, email, scheduling, shopping and document creation, but its branding and interface may limit its appeal against more assistant-like rivals.
Gemini Task is being offered free to users in the United States, Canada, Australia, India, Japan, and South Korea. The rollout is gradual, so availability may vary by account. Outside those markets, access requires a Pro or Ultra subscription, with notable exclusions including the UK and the European Economic Area.
The service is designed to receive assignments rather than just answer questions. Users can ask it to triage an inbox, draft replies, plan meals, add ingredients to Instacart, or prepare reports. That puts it in direct competition with newer consumer AI agents that are marketed as digital assistants rather than chatbots.
A standout capability is direct use through the phone’s native messaging app on iOS and Android. That allows users to send tasks as easily as texting a contact, including by voice through a phone assistant while driving. The format also creates a persistent thread that feels more like a real assistant relationship than a series of isolated prompts.
Because the agent sits inside Google’s ecosystem, it can create output directly in products such as Google Slides and potentially Google Sheets. In one example, it summarized reactions to a new OpenAI model and then turned the results into a presentation. That native integration gives Google an immediate productivity advantage without requiring extra setup.
Users can forward emails to the agent or copy it into a thread, then assign follow-up actions such as finding availability, locating a report or tracking updates. The system recognizes the sender’s identity through their email account and can reply with status updates. This turns routine email handling into an asynchronous workflow rather than a live chat session.
Google is preparing phone-call support so the agent can contact businesses, book appointments and place results on a Google Calendar. Payments through Google Pay are also planned, allowing purchases with user approval. Those features would move the tool beyond drafting and planning into direct real-world action.
The agent can follow up on its own after a failed task. In one test, after being unable to place a phone call, it later checked a business website, found no holiday schedule and then proposed sending an email inquiry instead. That kind of initiative is a key marker of emerging agent behavior rather than simple command response.
Connected services currently include apps such as Instacart, OpenTable, Peloton, SeatGeek, and Shopify. Users can also enable memory and other personal settings to improve relevance over time. Even so, the external app lineup remains narrower than some competing agents, which may matter for users who want broader automation.
The product is tucked inside Gemini under the plain name Task, with no separate identity, avatar or customizable personality. Regular chats and assigned tasks are mixed together in the same interface. That may make the experience feel like a feature rather than a dedicated assistant, especially compared with rivals that emphasize a named, persistent persona.
The system can create images by text request and generate dashboards, including recurring price trackers for products such as AirPods. It also appears to orchestrate subtasks by opening new internal task threads when needed. However, reliability remains uneven: some longer jobs are acknowledged but not clearly reported as finished until the user asks for an update.
Gemini Task gives Google a credible personal AI agent with strong built-in productivity tools and unusually convenient messaging-based control. Its long-term success may depend less on raw capability than on whether Google turns it from a hidden feature into a distinct assistant people actually want to rely on every day.
Ask a question