Google LLC’s Android team today announced it’s employing deeper integration of generative artificial intelligence and agentic AI to help mobile developers and publishers connect with customers and users. 

The company said at its annual Google I/O 2026 developer conference that said it’s expanding how users can find apps in the Google Play store on Android and the web using Gemini. This manifests as Ask Play within the store, a new AI-powered overlay experience with Gemini under the hood that turns app discovery into a natural conversation. 

The new experience can access the full context of different app experiences and listen to what the user wants through an iterative discussion to help them find what they want in the store. More and more users are turning to AI chatbots for discovery, and with Ask Play highlights, users can get a high-level summary of searches directly within search results to help them decide what apps or games they want from the ever-growing catalog. 

Google also announced updates to Engage SDK, which helps bridge and surface content between devices. For example, if users spend time watching a TV show on their Google TV and then walk away, they might notice a Play listing showing them that they can continue watching on their Android tablet when they’re away from home and have free time. 

Starting next month, existing users will see content notices directly in store listings. New tablet surfaces, including home screen Collections, are coming soon. Engage SDK surfaces can now scale across more than 80 different Play markets. 

Agentic AI expands across Android development

To aid developers in building the next generation of Android apps, Google showcased new capabilities for the Android command line interface, or CLI, now in stable version 1.0, that integrate with powerful agentic AI capabilities. 

This is a tool that, instead of running as a window on a screen, runs at a lower level using a text interface, which is closer to where most large language models operate. It also means that AI agents can “talk” to the CLI in its own language and get work done, allowing users to hook up whatever AI coding tool they want. 

Developers can quickly get up and running with an Android plugin package for Google’s Antigravity agentic coding environment, which now includes the Android CLI and skills. This provides Antigravity with tools and capabilities to use the CLI to perform the core tasks for mobile app development. 

To add more options, the updated CLI version introduced a new “android studio” command that allows the agent of developer choice to use the contextual capabilities of Android Studio to understand and perform actions. By running Android Studio alongside an AI agent using Android CLI, the agent can more readily navigate an Android codebase and better understand the ins and outs of projects. 

Native Android app creation

Google AI Studio, the company’s browser-based prototyping environment for developers powered by Gemini, can now build entire high-quality native Android apps in minutes. 

The new update brings the capabilities originally introduced through Gemini in Android Studio directly into the web-based environment. 

Native Android apps are built primarily in Kotlin and Jetpack Compose and are often iterated on using desktop editors such as JetBrains, not normally web-based environments. However, with this update, developers can launch a web browser, describe what they want and have the agent go to work in the cloud. 

Using a Google Play developer account, it’s also possible to publish directly from AI Studio for testing. The agent will automatically create the app record, package its bundle and upload it to an internal testing track in Google Play Developer Console. 

The app will become available to install within minutes, and it can be iterated upon within AI Studio. 

Moving beyond the browser is also simple. If the developer wants to break out of the web, AI Studio can wrap the project up into a ZIP format or export it directly to GitHub. This will allow developers to readily import anything built in the cloud into a different AI coding tool or development environment. 

 Image: SiliconANGLE/Microsoft Designer