Gemini is getting its first agentic capabilities

Google's Gemini AI is introducing its first agentic capabilities through task automation, starting with select Pixel 10 phones and the Samsung Galaxy S26 series. This feature allows Gemini to execute multi-step tasks within apps, such as hailing an Uber or placing a DoorDash order, by launching the app in a virtual window and navigating the steps. Users can monitor the process and intervene if necessary, with Gemini notifying them for critical decisions or when items are out of stock. According to Android ecosystem president Sameer Samat, this is part of Google's vision for Android as an "intelligence system," and this automation ability will extend to the next major Android release. Gemini achieves this through reasoning, utilizing existing app functions frameworks, or by figuring out the steps itself. While currently limited to a few apps like Uber and Grubhub for an early preview in the US and Korea, Google is working with developers to embrace this new capability.

(Source：The Verge)