Imagine if your weekly grocery shop was as simple as pressing one button? The word “agentic” has been thrown around a lot this past year, but AI-driven commerce experiences that are truly agentic — meaning AI agents handle multiple steps in an order, even the whole thing, without input from a user — have been slow to arrive. These kids of automated experiences may be just around the corner though.
Uber, DoorDash and a number of other companies in the food, grocery and rideshare categories are participating in early tests of task automation on Google’s Gemini AI platform that will execute multi-step processes like booking a ride or reordering a meal in popular mobile apps.
The capability is launching in beta in the Gemini app for Pixel 10, Pixel 10 Pro and Samsung Galaxy S26 series phones, initially in the U.S. and Korea. Users who meet those requirements simply long-press the power button and ask Gemini to help book a ride home or reorder on DoorDash. Automation will begin with the user’s request and end as soon as the task is completed, and users can monitor Gemini’s progress at any time via live notifications. If things aren’t going to plan, users also can jump in to modify or change what Gemini is doing.
