Google brings Gemini Intelligence to Android Auto — for now, Pixel only

For years, the in-car assistant has promised the same thing: "tell me what you need and I'll do it." It worked poorly, because the assistant knew nothing about your life. Now Google is trying to close that gap: Gemini Intelligence gets access to data from your phone's apps and shows the result right on the car's screen.

What exactly Google is launching

Gemini Intelligence is moving beyond the Pixel 11 lineup. The feature is coming to the Pixel 10, the upcoming Motorola Signature 27, and several foldable and flagship Samsung Galaxy models. But in Android Auto, only eligible Pixel phones will get it. It's a familiar Google move: some new features go first to its "own" phones.

In the car, the feature gathers information from apps and displays it on the infotainment screen. An example from Google's blog: a friend texts you "where are we meeting for Pilates?" Gemini finds the booking confirmation email in your inbox and offers to send it with a single tap. No need to dig through your conversations while driving.

A ZDNET reviewer adds similar scenarios: "remind me to buy Mom a card before her birthday" or "what's the fastest route to the dentist." For tasks like these, the assistant needs context from messages, email, and the calendar.

Two levels of data access

The most interesting part is how the scheme itself is built. Google now has two tiers. Personal Intelligence works on Android phones without Gemini Intelligence and connects Gemini to Google Maps, Photos, and Gmail. You can ask it to route you to a restaurant with a particular cuisine or to find, in your photos, the car wash you snapped last week. All data comes from Google services.

Gemini Intelligence goes deeper: it relies on the phone's own data, including messages, and behaves more proactively. The line between the versions is drawn not by how "smart" the model is, but by which data it is allowed to touch.

What is still unknown

The main expectation concerns proactivity. The ZDNET author uses Siri in CarPlay, which tells them when it's time to leave for dinner, a flight, or a doctor's appointment, taking traffic into account. According to the author, Gemini Intelligence in Android Auto seems likely to do the same. But this is unconfirmed: the journalist has requested comment from Google and promises to update the piece. For now, there is only the example with the email and Pilates.

There are also practical limitations. You need the Gemini app and additional setup. The feature is available in selected countries and to users 18 and older, and the set of capabilities will vary. Separately, a reminder: Android Auto now requires Android 11 or newer, so it's worth checking compatibility.

What to do with this

For product managers and developers, the takeaway isn't about Pixel. The in-car assistant is becoming a layer on top of apps: it pulls confirmations, emails, and events out of what's already on the phone. If your product sends bookings, orders, or tickets in a clear form, the assistant will find them. If not, your product simply doesn't exist for it. It's worth looking at how your emails and notifications appear through the eyes of such an agent.

The second lesson concerns tiered access. Google separates "data from our services" from "everything on the phone, including messages." This lowers the trust threshold: the user first gets a safe version and only then decides whether to let the agent go deeper. If you're designing your own agent, think about which features can be offered without sensitive data.

The third observation concerns the interface. In a car, minimal action matters, so "one tap" instead of a dialogue is a good benchmark for any in-car scenario.

A caveat: this is a consumer-outlet overview, not a technical breakdown. The piece says nothing about an API for third-party developers, about exactly which data goes to the cloud, or about timing by region. Those details will have to come from Google.