Google is testing a new feature that lets its Gemini AI assistant call businesses for users. The feature is called "Call for Me."
The test is starting small. It will first be available to Pixel 11 owners in the U.S. who pay for a Gemini subscription.
Users will also need the beta version of Google's Phone app for Android. Google describes the feature as an experiment for now.
The launch comes as other AI agents, such as Meta's Muse and Instinct, can already make calls for users.
What the Feature Can Do
Google has offered AI-guided calling tools before, but those versions had limits. The new feature lets Gemini share personal information that the user approves during a call.
Google says this allows the AI to take a wider range of actions for the user.
Examples include calling a store to check if a product is in stock or making a restaurant reservation. Gemini can also move an appointment to a new date or place items on hold.
The AI can introduce itself when a business answers. It can navigate automated phone menus, wait on hold, and handle the conversation with the person on the other end.
Calls are made from the user's own phone number and dialed directly from their phone. Users can follow along through a live transcript of the conversation.
Google says users can take over the call at any time.
Years of Calling Experiments
The launch builds on years of work on automated calling at Google.
In 2018, at its I/O developer conference, Google showed Google Assistant calling a salon to book an appointment. The AI added small sounds like "umm" and "ahh" to make the call sound natural.
Last year, Google released "Ask for Me." That feature let Gemini call businesses to ask about pricing or services.
Pixel phone owners have also had access to "Hold for Me" and "Talk to a Live Rep." These tools navigate phone trees and wait on hold until a real person comes on the line.
Another tool, "Direct My Call," shows the menu numbers users need when dealing with automated phone systems.
Google said it is starting the new test at a small scale because "real-world conversations are nuanced."
The company said it needs time to get the feature right before rolling it out more broadly.
