Google has been promising that AI assistants would eventually just do things on your phone - not answer questions about doing things, but actually tap buttons, fill forms, and complete tasks inside real apps. With Gemini's new task automation on the Pixel 10 Pro and Galaxy S26 Ultra, that future has technically arrived. The word "technically" is doing a lot of heavy lifting.
The feature lets Gemini take control of your phone and operate apps on your behalf. Right now the supported list is short: a handful of food delivery and rideshare services like Uber and DoorDash. You tell Gemini what you want, and it navigates the app, taps through menus, and places orders or hails rides for you.
What It Actually Looks Like in Practice
Slow. You can watch Gemini methodically tap through screens at a pace well below what your own thumbs could manage. It pauses, processes, taps, waits, processes again. Ordering a burrito through DoorDash yourself takes maybe 90 seconds. Watching Gemini do it feels like supervising someone use a smartphone for the first time.
But here's the thing that makes this more than a party trick: it works. Gemini reads app interfaces it wasn't specifically trained on, figures out which buttons to press, handles the back-and-forth of selecting options and confirming choices, and gets to the finish line. That's a genuinely hard computer science problem - understanding arbitrary user interfaces in real time and interacting with them correctly.
The Gap Between Demo and Daily Driver
No one is going to use this to save time right now. The app support is too narrow and the execution too slow. But the underlying capability - an AI that can see your screen, understand what's on it, and take action across any app - is the foundation for something much more useful once speed and reliability catch up.
Google isn't alone here. Apple Intelligence has been creeping toward similar on-device automation, and OpenAI's Operator tackles the same idea from the desktop browser side. The race to build an AI that can actually use software rather than just talk about it is one of the more consequential product battles in tech right now.
For the moment, Gemini's task automation is a proof of concept you can actually try on your phone. It proves the concept convincingly. It just doesn't yet prove it's faster than doing things yourself.