Google Is Testing a Feature That Lets Gemini Make Business Calls for You
Author : Ravit Jain
Google is testing a new capability within its Gemini AI assistant that allows the system to place phone calls to businesses on behalf of users. The feature, named "Call for Me," represents a significant step forward in AI-driven voice interaction, and it puts Google alongside other technology companies that have been pushing autonomous AI agents into real-world communication scenarios.
The rollout will begin on a limited basis, initially available only to Pixel 11 owners in the United States who hold an active Gemini subscription. Users will also need to be running the beta version of Google's Phone app for Android, as the feature remains in experimental territory for now.
What "Call for Me" Can Do
While Google has offered AI-assisted calling tools in the past, those earlier implementations were relatively narrow in scope. The new "Call for Me" feature is designed to handle a broader range of tasks by allowing Gemini to share user-approved personal information during calls, enabling it to take more meaningful action on the caller's behalf.
According to Google, the AI is capable of managing genuinely complex interactions. It can contact a retailer to check whether a specific item is available in store, make a restaurant reservation, reschedule an existing appointment, or request that a product be placed on hold. The system can introduce itself to whoever answers, navigate automated phone menus, wait on hold if necessary, and then carry on a conversation with a representative once one becomes available.
Importantly, users retain control throughout the process. The call is placed directly from the user's own phone number, and a live transcript of the conversation is displayed in real time. At any point during the call, the user can choose to step in and take over the conversation themselves.
Building on Years of AI Calling Experiments
This launch is not Google's first foray into AI-powered phone calls. The company made headlines several years ago when it demonstrated Google Duplex at its annual I/O developer conference, showing off Google Assistant placing a call to a hair salon to book an appointment. That demonstration drew widespread attention partly because the AI incorporated natural-sounding filler phrases - such as "umm" and "ahh" - to make the interaction feel more like a genuine human conversation.
More recently, Google introduced "Ask for Me," a feature that allowed Gemini to reach out to businesses on a user's behalf for informational purposes, such as asking about service availability or pricing. Pixel device owners have also had access to a set of related tools: "Hold for Me" and "Talk to a Live Rep" both allow the phone to navigate automated menus and wait on hold until a human agent is available. A separate option, "Direct My Call," displays the keypad inputs needed to move through interactive voice response systems, giving users a faster path through phone trees.
A Cautious Rollout Strategy
Google has been deliberate about how it is introducing "Call for Me" to the public. The company acknowledged that it is starting at a small scale precisely because live phone conversations are inherently unpredictable and context-dependent. Real-world exchanges between people - and between AI systems and people - involve nuance that is difficult to fully anticipate or engineer around in advance.
By limiting the initial release to a small group of users, Google says it intends to gather the kind of real-world feedback needed to refine the feature before making it more widely available. The experimental nature of the rollout signals that the company views this as an ongoing process of iteration rather than a finished product ready for mass deployment.
The development also comes as other technology companies accelerate their own efforts in autonomous AI agents. Meta, for instance, has been advancing AI systems capable of acting on behalf of users in various contexts, reflecting a broader industry trend toward AI that does not just answer questions but actively carries out tasks in the real world.
Source Credits : TechCrunch


