Google has begun testing a new AI‑driven feature that lets its Gemini model place phone calls to local businesses on behalf of Pixel 11 owners. Branded “Call for Me,” the early experiment allows users to delegate routine tasks—such as reserving a table, checking stock, or rescheduling an appointment—while watching a live transcript of the conversation and taking control at any moment.
What happenedThe rollout is limited to Pixel 11 devices in the United States and requires a paid Gemini subscription as well as the beta version of Google’s Phone app for Android. Unlike earlier AI‑assisted dialing tools, Gemini now uses the user’s own phone number to place calls, dialing directly from the device. The system can introduce itself, navigate automated menus, wait on hold, and conduct the full dialogue with the business.
Key capabilities highlighted by Google include:
- Real‑time sharing of approved personal information during the call, enabling more complex interactions.
- Live transcription that lets the user follow the exchange as it happens.
- A “take‑over” button that instantly transfers the call to the user if they wish to intervene.
- Ability to handle tasks such as confirming product availability, making restaurant reservations, moving appointments, or placing items on hold.
The feature builds on a series of prior experiments. Google previously demonstrated the Google Assistant making a salon reservation at its I/O conference, and last year introduced “Ask for Me,” which let Gemini inquire about pricing or services. Other Pixel‑specific tools—“Hold for Me,” “Talk to a Live Rep,” and “Direct My Call”—have helped users navigate IVR (interactive voice response) systems, but none offered the full conversational hand‑off that Gemini now provides.
Google notes that the experiment is being launched at a small scale because “real‑world conversations are nuanced,” and the company wants to refine the experience before a broader release.
Why it mattersIf successful, the feature could reshape how consumers interact with businesses over the phone. By automating routine calls, Gemini may free up time for users and reduce the friction of waiting on hold—a common pain point in customer service. The live transcript and instant takeover option address privacy and control concerns, giving users visibility into what the AI says and the ability to intervene instantly.
From a business perspective, the technology could lower call‑center volume for simple inquiries, allowing human agents to focus on more complex issues. For Google, the experiment extends the utility of its Gemini subscription beyond text‑based queries, positioning the model as a multi‑modal assistant that can act in the physical world through voice and telephony.
The bigger pictureGemini’s new calling ability arrives amid a growing field of AI agents that can make phone calls on a user’s behalf. Meta’s Muse and Instinct have already demonstrated similar functionality, signaling a broader industry push toward conversational agents that can act autonomously in real‑time interactions.
Google’s approach differs in several ways. First, the calls are placed from the user’s own number, preserving the appearance of a personal outreach rather than a generic AI line. Second, the integration with the native Phone app and the requirement for a Gemini subscription tie the feature tightly to Google’s ecosystem of services and hardware, particularly the Pixel line.
The experiment also reflects Google’s incremental strategy. Over the past few years the company has layered increasingly sophisticated voice‑based tools—starting with simple “Ask for Me” queries, then adding hold‑management features, and now full conversational hand‑off. Each step has been tested in limited releases, allowing Google to gather real‑world data on user behavior, privacy implications, and technical reliability before scaling.
What happens nextGoogle has indicated that the current rollout is a small‑scale test and that further expansion will depend on how the feature performs in real‑world usage. The company emphasizes the need to fine‑tune the system because of the nuanced nature of live conversations. Users who enroll can watch the live transcript, intervene at any point, and provide feedback through the beta Phone app.
The next steps will likely involve:
- Collecting data on call success rates, user satisfaction, and privacy compliance.
- Refining the AI’s ability to handle varied business scripts and unexpected dialogue turns.
- Expanding the feature set based on user feedback, possibly adding scheduling integrations or multi‑step task automation.
If the trial proves effective, Gemini could become a standard part of Google’s subscription offering, further blurring the line between digital assistants and everyday personal tasks.
The information in this article is drawn from reporting by The Verge and TechCrunch on September 24, 2026.
