Google's Gemini Intelligence adds agentic AI and custom widgets to Android
Copy a grocery list to shopping cart with a single press – Google's AI now completes tasks across apps.
Google unveiled a suite of Gemini Intelligence AI features at its Android Show: I/O Edition, building on agentic capabilities first previewed at the Samsung Galaxy S26 launch. The assistant can now execute multistep tasks across apps – for example, copying a grocery list from Notes and adding items to a shopping cart, with the phone screen providing context. Users trigger this by pressing the power button and describing the task. The auto-browse feature, which allows Gemini to surf the web and complete actions like booking appointments, is also coming to Android. Additionally, Gemini in Chrome will arrive in late June, summarizing web pages and answering questions.
Other updates include form filling via Personal Intelligence (opt-in), and Rambler in Gboard – a dictation feature that transcribes speech and removes filler words while preserving the user's tone. Most notably, Google introduced the ability to vibe-code custom Android widgets by describing them in natural language (e.g., "Suggest three high-protein meal prep recipes every week"), with outputs following Material 3 design. These AI features will debut on latest Samsung Galaxy and Google Pixel devices this summer, rolling out to other Android devices later in the year.
- Gemini can execute multistep tasks like copying grocery lists from Notes to shopping apps, with final confirmation required.
- Auto-browse lets Gemini surf the web and complete tasks (e.g., booking appointments) on Android; Gemini in Chrome arrives in late June.
- Users can vibe-code custom Android widgets using natural language (e.g., 'Suggest three high-protein recipes weekly') with Material 3 design.
Why It Matters
Google’s agentic AI transforms Android into a proactive assistant, reducing manual app switching and enabling custom UI widgets.