Gemini’s task automation is here and it’s wild
The new feature lets your phone use apps itself, starting with Uber and Starbucks.
Google and Samsung have launched a beta for Gemini's long-promised task automation feature, with the Galaxy S26 Ultra as one of the first devices to get it. This represents a significant leap from conversational AI to actionable AI, where the assistant can directly interact with apps on a user's behalf. Starting with partners like Uber and Starbucks, Gemini can now execute multi-step workflows—such as ordering a ride or food—within a secure virtual window, performing clicks and scrolls autonomously after receiving a simple text prompt.
In early testing by The Verge, the system successfully handled a request to "order an Uber to the airport," asking clarifying questions and navigating the app's interface to input the destination. For a more complex Starbucks order, Gemini autonomously scrolled through the menu to find a flat white and correctly specified that a chocolate croissant should be warmed. Crucially, the automation stops before any final payment or confirmation step, requiring user review. This feature, which users can watch in real-time and interrupt at any point, moves AI assistants closer to becoming true digital agents that can execute tasks, not just answer questions.
- Gemini's new automation can use apps like Uber and Starbucks in a virtual window to complete tasks.
- The AI performs multi-step actions like inputting destinations and scrolling menus, but requires final user confirmation.
- The feature is in beta, first rolling out on the Samsung Galaxy S26 Ultra, marking a shift from conversational to actionable AI.
Why It Matters
This transforms AI from an information tool into an action-taking agent, potentially saving users significant time on routine digital tasks.