Google is giving Gemini a new job: making phone calls so users do not have to.
The company has started testing Gemini Call for Me, an AI-powered feature that can contact businesses on a user’s behalf, ask questions, book appointments and even wait through phone menus. Instead of simply placing the call, Gemini can handle the conversation before returning with a summary of what happened.
The feature is currently limited to eligible Pixel 11 users in the US, but it gives a clearer picture of where Google wants its AI assistant to go next. Gemini is moving from answering questions on a screen towards completing everyday tasks that normally require direct human interaction.
What Can Gemini Call for Me Actually Do?
The new feature is designed around ordinary phone calls that people frequently postpone or find inconvenient.
A user can ask Gemini to contact a restaurant and reserve a table, phone a local store to check whether a product is available or call a service provider for a price quote.
It can also contact businesses to confirm appointments and work through automated phone menus.
The main difference from a standard voice assistant is that Gemini AI calling can continue the interaction itself rather than simply dialling a number and handing control back to the user.
Once the task is complete, Gemini provides a summary of the outcome inside the app.
That could make the feature particularly useful for relatively simple calls where the user knows exactly what information or result they need.
Users Start by Giving Gemini Clear Instructions
The process begins inside the Gemini mobile app.
Users can tell Gemini what they want done in ordinary language or explicitly use the “@Call for me” command.
The instruction can include details such as the business name, preferred date or time and any alternatives Gemini should accept if the first option is unavailable.
Before making the call, Gemini presents a summary of the task.
The user can check which phone number will be called and review the personal information Gemini plans to share, including their name and contact details.
Nothing begins until the user approves the calling task.
That confirmation stage is important because the AI may need to share information with a business in order to complete a reservation or appointment.
Gemini Identifies Itself as AI on the Call
Google is not trying to make businesses believe they are speaking with a human customer.
At the beginning of each call, Gemini Call for Me identifies itself as a Google AI assistant calling on behalf of the user.
It also explains that the conversation is being recorded and states the user’s name.
Businesses therefore know they are dealing with an automated system before the conversation continues.
Google already has experience with automated business calls through earlier Search and Maps features, but the Gemini version gives users much more direct control over the task.
Instead of Google calling businesses mainly to confirm information, the AI is acting on a specific customer’s behalf.
You Can Listen In or Take Over the Conversation
Users are not completely removed once Gemini starts speaking.
The app provides a live transcript of the call, making it possible to follow the conversation without participating directly.
Users can also listen to the live audio.
If the conversation moves in an unexpected direction or the business asks something the AI cannot answer, the user can tap a button and take over.
Gemini then announces that the user is joining and leaves the conversation.
This gives Google Gemini a useful middle ground between full automation and a traditional phone call.
People can let the AI handle routine sections while still having the option to step in if human judgement becomes necessary.
Gemini Can Even Wait on Hold
Anyone who has called a large business knows that actually speaking to someone may be the easiest part. Waiting through menus and hold music can take much longer.
Google is extending Call for Me to handle some of that work too.
Gemini can navigate phone menus and wait on hold, returning control when necessary.
That builds on technology Google has already used through features such as Hold for Me and Direct My Call on Pixel devices.
The difference is that the AI can now continue beyond the waiting process and attempt to complete the actual task.
For simple service calls, that could save users several minutes without requiring them to keep a phone at their ear.
There Are Strict Limits on What Gemini Can Do
Google is putting several restrictions around Gemini AI calling.
The assistant cannot make emergency calls, including calls to 911.
It cannot complete financial transactions or provide sensitive financial information such as credit card details over the phone.
Gemini is also restricted from sharing sensitive personal information such as passwords and Social Security numbers.
Telemarketing is not supported either, meaning the feature cannot be used to make automated sales calls.
Users are also subject to daily limits on how many calls the AI can make.
These restrictions show that Google is starting with lower-risk everyday tasks rather than allowing Gemini to handle any kind of phone conversation.
Availability Is Very Limited for Now
Most users will not have access immediately.
Google is gradually rolling out the feature to people aged 18 or older in the United States who own a Pixel 11 and use a US SIM card.
A Google AI subscription is required.
Users also need the public beta version of the Phone by Google app, the latest Gemini app and English set as the device language.
For now, Gemini can only make calls to US phone numbers.
That makes Gemini Call for Me more of an early experiment than a mainstream Android feature.
Google will be able to see how the system behaves during real conversations before deciding how widely it should be released.
AI Assistants Are Moving From Answers to Actions
The more interesting part of Call for Me is what it says about the direction of AI assistants.
For several years, systems such as Gemini have mainly been judged by how well they answer questions, write text or generate images.
The next phase is increasingly about agents that can do something outside the chat window.
A phone call is a useful test because real conversations can be unpredictable. A restaurant may suggest a different time, a shop may offer another product or an automated menu may change unexpectedly.
Handling those situations requires more than generating a prewritten response.
If Google Gemini can reliably complete routine calls, the technology may eventually become useful for a much broader range of everyday tasks.
For now, the rollout is narrow and experimental. But the basic idea is significant: instead of asking Gemini what number to call, some users can now simply tell it what they need and let the AI make the call itself.
Source:
The Indian Express: Google’s Gemini AI assistant can now call businesses for you: Here’s how it works
