Discover Brum
Membership Login

Google tests Gemini calling businesses for you

Call for Me can book appointments and wait on hold on a user's behalf.

Discover Brum Newsdesk

2 min read

A smartphone screen
A smartphone screen. Photo: Skitterphoto / Wikimedia Commons, CC0

Google has begun testing a feature that lets its Gemini assistant phone businesses on a user's behalf, to book appointments, make reservations or check whether a shop has something in stock.

The feature, called "Call for Me", started testing on 24 September.

How it works

Users can ask Gemini to call a business to place an order, schedule an appointment, make a reservation or ask whether a product is available. The assistant can find its way through automated phone menus and wait on hold until a person answers.

The call is made from the user's own phone number, and a live transcript is shown on screen so the user can take over at any point. Gemini can share personal details that the user approves, allowing it to do more on the call.

A small start

The early preview is available to paying Gemini subscribers in the United States who have Google's Pixel 11 phones, and it requires the beta version of Google's Phone app for Android.

Google said it was starting small because "real-world conversations are nuanced," and it wants time to get the feature right before offering it more widely.

An idea with a history

Google first showed a system that could make phone calls for people in 2018, when its Duplex technology booked a hair appointment and a restaurant table in a demonstration that surprised audiences with how human it sounded. That raised questions about whether people should be told when they are talking to a machine.

The new feature builds on that work using Gemini, Google's current family of AI models.

Part of a bigger shift

Call for Me is one of a growing number of features in which AI assistants act for their users, rather than only providing information. Google is also testing shopping directly inside Gemini in India. Such agents promise to save time, but also raise questions about how businesses will handle calls from machines.

All