Google has announced Gemini Intelligence, a set of AI capabilities for Android that lets the Gemini assistant carry out tasks inside apps on a user’s behalf. The company presented the system at its Android Show event on May 12, describing a shift from an assistant that answers questions to an agent that can complete multistep jobs across different apps.
Google says Gemini Intelligence will arrive first on Google Pixel and Samsung Galaxy devices this summer, with other Android phones to follow.
An agent that works across apps
According to Google, Gemini Intelligence can understand what is on the screen and take actions within apps when a user asks. The examples the company gave include ordering groceries, filling in forms, moving information from one app to another and handling tasks that involve several steps. The assistant is meant to follow the user’s instructions rather than operate entirely on its own, but it reduces the amount of tapping and switching between apps that those tasks normally require.
The idea of an AI that uses apps for people, sometimes called agentic AI, has been a focus for the major technology companies over the past year. Google’s approach integrates the capability into the operating system, which gives Gemini access to app content and actions in a way that a standalone chatbot app does not have.
Google also used the event to show other Gemini-driven features, including generative widgets that the assistant can build on request, according to coverage of the announcement.
Googlebook laptops join the lineup
Alongside the phone features, Google introduced Googlebook, a new category of laptops designed around Gemini from the start. The devices run an Android-based operating system that also supports Chrome apps. Google named Acer, Asus, Dell, HP and Lenovo as hardware partners, with the first devices expected in the fall of 2026.
Taken together, the announcements show Google trying to make Gemini the common layer across phones and computers, rather than a product confined to a search box or a chat window.
Competitive context
Google’s timing puts pressure on Apple, which has been working to deliver a more capable, Gemini-powered Siri, and on Microsoft, whose Copilot features are built into Windows. Each company is now competing to be the first to put a reliable personal agent in users’ hands.
How well the system performs in everyday use remains to be seen. Agents that act inside apps raise practical questions about accuracy, permissions and what happens when a task goes wrong, and Google has not yet published detailed documentation on how users will review or undo actions the assistant takes. Those details should become clearer when the features begin rolling out to Pixel and Galaxy phones.
