The Pulse

Meta’s Muse can call US businesses for paid subscribers

Meta’s Muse AI agent can now place calls to U.S. businesses for paid subscribers, including calls about appointments and reservations. The rollout follows an earlier test that involved human callers and drew concerns from employees about privacy and disclosure.

AI.info Team · · fact-checked · against 5 sources ↓ · Published 13 Nov 2026

Paid Muse subscribers in the United States can now ask Meta’s AI agent to place calls to businesses on their behalf. The expansion follows a limited test of the calling feature that began in September.

Muse announced the feature on X: “Let Muse handle the wait times while you get back some free time. Introducing outbound calling, available now for paid US subscribers. ☎️” Alexandr Wang, Meta’s chief AI officer and head of its Superintelligence Labs division, also announced the availability on November 12. Previously, a limited group of early adopters could try it after asking Muse for access, TechCrunch reported.

What subscribers can request

Wang suggested uses including calls about insurance or cable service, contacting the DMV and making dinner reservations. TechCrunch reported that the feature could be used for restaurants that do not offer online reservations.

Muse’s paid plans start at $20 per month for 500 million Muse tokens per week. A higher-priced plan costs $100 per month for 3 billion tokens, according to TechCrunch.

Muse launched on September 8 as a personal AI assistant. Meta said last month that it had more than 10 million users, TechCrunch reported.

The earlier human-caller test

Before the wider paid rollout, Meta also tested having human contractors handle some calls made through Muse. Reuters reported in September that the test followed reports from some Muse users that businesses often hung up after recognizing an AI caller, according to PYMNTS.

Internal company posts showed that Meta later ended the human-caller test for the time being, PYMNTS reported, citing Reuters. Employees raised concerns that sensitive user information could reach contractors and that the test could undermine security measures Meta promoted at Muse’s launch, the report said.

Internal posts seen by 404 Media described a test in which Muse could hand a request to a trained human agent, who would place the call. One tester was not told a human had made a call until afterward, an employee wrote in communications reviewed by the outlet.

Meta spokesperson Daniel Roberts told Reuters that the company ran the test to gather feedback before releasing features publicly and described employee response as “overwhelmingly positive,” according to PYMNTS. Roberts also said the company was working with merchants on the potential calling feature and would release it when ready and with proper disclosures.

Calls in practice

In an example reported by Engadget, Meta Superintelligence employee Matt Schlicht shared a recreation of a call his agent made to a Volvo dealership to ask about an appointment. Schlicht said Muse waited on the line for 15 minutes.

Engadget reported that the agent did not explicitly identify itself as AI in the shared call, though it said it was calling on Schlicht’s behalf and transcribing the conversation. Other users’ experiences with the earlier feature varied: 404 Media reported that some calls worked, while others failed or ended when the recipient hung up.

Calling is not Muse’s only recent communication feature. Wang also said the option to create a Muse email address would soon reach free users, TechCrunch reported.