Gemini Intelligence
The new features will first appear on recent Samsung Galaxy devices and Google Pixel smartphones, with more Android devices to follow later this year.
A key part of Gemini Intelligence is automation between apps. Android should soon be able to better understand what you want to do and perform multiple actions in succession. Think of automatically booking a spin class, adding textbooks to a shopping basket or arranging a taxi. In doing so, Gemini also uses information visible on your screen or in images.
For example, Google shows how to convert a shopping list from a notes app into an online shopping basket, simply by asking Gemini to do so. Even a picture of a travel brochure can be enough to automatically look up similar holidays.
In Chrome
In addition, Google Chrome is getting a smarter AI assistant. Gemini in Chrome can summarise web pages, compare information and even perform certain tasks, such as booking appointments or reserving parking spaces.
Autocomplete also gets smarter. Android can use data from linked apps with Gemini to fill in forms faster. According to Google, this remains optional and you keep control over which data can be used.
Also notable is the arrival of Rambler, a new feature in Gboard. It lets you talk freely while Gemini automatically converts your spoken text into a neat and readable message. The AI extracts stop words, repetitions and hesitations, making messages look more natural. Rambler also supports multiple languages interchangeably.

Furthermore, Google is experimenting with AI widgets. Via "Create My Widget", you will soon be able to create widgets by simply describing what you want to see. For example, a widget with protein-rich recipes or a compact weather dashboard for cyclists.
Visually, too, Android is changing with it. Gemini Intelligence uses an updated version of Material 3 Expressive, with smoother animations and a quieter design that should help focus better. With Gemini Intelligence, Android is increasingly shifting towards a system that actively thinks with you and takes tasks off your hands. How well that works in practice will have to be seen later this year when the first features become more widely available.