Google Gemini on Android 16 phones to perform in-app tasks on users' behalf

Reportedly, Android 16 will feature an advanced version of Google Gemini, which will enable it to perform in-app tasks on the user's behalf such as ordering food

Google is reportedly planning to enhance its Gemini AI into a fully-fledged Agentic AI assistant for smartphones with Android 16. According to a report by Android Authority, the Android 16 update will introduce new "app functions" that will enable Gemini AI to take actions within apps on behalf of the user.

 

The report mentions that developer documentation released with the first developer preview of Android 16 earlier this month highlights a new "app functions" feature. Google defines app functions as a "specific piece of functionality that an app offers to the system," adding that "these functionalities can be integrated into various system features." The document outlines how Google plans to use these app functions, with examples such as a food ordering app implementing a function to place orders. Gemini AI would then be able to execute this action without requiring the user to open the app.

 

While details remain scarce, these app functions could allow Gemini AI to perform a wide range of actions within third-party apps. Currently, Gemini AI can only perform limited in-app actions through Gemini Extensions, which grant access to functionalities in native Google apps and select third-party apps like WhatsApp. The new Utilities extension introduces a few features, such as the ability to open specific apps and websites and adjust display and volume levels. However, Gemini Assistant is unable to carry out advanced multi-step functions without human input.

Additionally, Google is reportedly developing a large action model aimed at automating web-based tasks for users. Last month, The Information reported that Google is working on an AI agent codenamed “Project Jarvis,” designed to execute tasks on the web, such as booking tickets or conducting research. This AI agent is expected to be powered by a future version of the Gemini AI model and will be specifically tailored for use within the Google Chrome browser.