Gemini Adds Uber and Food Ordering Task Automation on Pixel 10 and Galaxy S26

Google's mobile AI roadmap is moving beyond chat-style assistance into direct app execution. According to The Verge, Gemini task automation is rolling out in early preview for select Pixel 10 and Samsung Galaxy S26 devices, starting with practical flows such as booking a ride and assembling a food order. The key product decision is not full autonomy at checkout: Gemini prepares and navigates the process, while the user still confirms and submits the final order.

How the workflow is designed

The interaction starts with a natural-language request. A user can ask Gemini to arrange a ride or put together a delivery order, and Gemini then opens the target app in a virtual window and progresses through each step. Users can watch in real time, interrupt, or take over at any point. If there are branching choices, inventory issues, or ambiguous options, Gemini asks for input before proceeding. This keeps the automation useful without removing user agency at critical transaction points.

What powers the automation

The Verge reports that Google is combining multiple technical paths instead of relying on a single integration model. In demos, Gemini 3 performs reasoning over UI flows, selecting options and handling alternatives. Developers can also expose actions through MCP or Android app functions, while Gemini can still attempt generalized UI-driven automation where explicit hooks are unavailable. Android ecosystem president Sameer Samat positioned this as part of a broader shift toward treating Android as an "intelligence system," and indicated these capabilities are tied to Android's next major release cycle.

Preview limits and strategic implications

The initial rollout is intentionally narrow. Availability starts in the US and Korea, and app coverage is limited to a small set that includes Uber and Grubhub, with demonstrations also showing food ordering scenarios such as DoorDash-style flows. Even with these constraints, the significance is substantial: mobile AI is transitioning from response generation to execution orchestration inside third-party apps.

For platform teams and app developers, this creates immediate priorities. Reliable action schemas, clear interruption handling, and secure confirmation boundaries become core requirements. For consumers, trust will depend on transparent progress, easy override controls, and clear final-approval checkpoints. In product terms, Google is testing a hybrid model where AI handles repetitive navigation while humans remain accountable for payment and commitment decisions.

If this preview expands successfully, task automation could become a default layer for everyday mobile actions such as transportation, commerce, and scheduling. That would make assistant quality less about answer style and more about dependable completion across real app ecosystems.