Google did what Apple couldn’t
Google is adding new “agentic” features to Gemini, allowing it to handle multi-step tasks on your phone.
This will launch first on the Pixel 10, Pixel 10 Pro, and Samsung Galaxy S26.
In a demo, Gemini read a group chat, figured out what everyone wanted to eat, and prepared a food order through a delivery app.
The user was then asked to review and confirm it. The demo was pre-recorded.
This is part of Google’s push to turn Gemini into a tool that can take action, not just respond.
It follows recent updates like auto-browsing in Chrome.
Here's what you should know:
Gemini is moving from chatbot to task assistant
Apple’s similar features are still delayed
Early performance will depend on app access and reliability
The “just handle it” era
The timing stands out.
Apple showed similar Siri features in 2024, including the ability to understand on-screen context and act across apps.
These features were later delayed and have still not been launched in 2026.
Google’s update marks progress in agent-based AI, but there are still open questions.
The feature is launching in beta, and performance may vary depending on app support and real-world use.
This only works if it doesn’t mess up the order. One wrong topping and it’s over. - MG