Meta's Muse Can Now Phone U.S. Businesses for Paying Subscribers
Meta's Muse AI agent can now place calls to U.S. businesses on behalf of its users, and only paid subscribers can use the feature. TechCrunch reported the change on November 12, 2026.
Meta first tested the business-calling feature in September 2026. During that phase, only a select group of early adopters could try it, and they had to ask Muse for access before they could use it. The wider release extends the feature to paying subscribers.
Muse subscriptions start at $20 per month, which includes 500 million Muse tokens per week. A higher tier costs $100 per month and includes 3 billion tokens. Tokens are the chunks of text a language model reads and writes, so the allowance works like a usage meter. Because access is tied to that token budget rather than a flat call count, a subscriber's calling capacity is bounded by the weekly or monthly allowance. TechCrunch lists both tiers.
Meta said in October 2026 that Muse had more than 10 million users.
The free tier is getting a different addition. A Muse email address feature will soon roll out to free users. The calling capability stays behind the paywall, which means the two most visible new functions split along the subscription line: free users will receive an address, while outbound calling belongs to paid plans.
For developers and operators, the more interesting question sits on the receiving end. An agent that dials a business on a user's behalf moves routine work, such as checking hours, confirming availability or asking about a booking, from the consumer onto the company's phone lines and staff. Whether that trade works out will depend on how reliably the agent handles automated phone menus, interruptions and ambiguous answers.
The broader context here is a shift in who initiates a transaction. For two decades, businesses built web forms, chatbots and call-center IVR systems (the automated "press 1 for billing" menus) to absorb inbound demand. A consumer-side agent that places the call reverses that flow, and it does so over the same public telephone network businesses already staff. Companies that see a sharp rise in agent-originated calls will need a way to separate them from scams and robocalls, and the usual caller-identification and consent frameworks were not designed with an AI intermediary in mind.
In my view, the upside is real if the tool works as advertised. Someone stuck on hold during a lunch break, or someone who needs three repair shops checked before choosing one, could hand that errand to software and get the answer back. The downside is just as concrete: a flood of automated calls to small businesses that lack the staff to screen them could erode the phone's value as a channel. How Meta identifies its agent, and what businesses can do to opt out or filter it, will shape which outcome dominates.
For now, the facts are narrow. The feature is live for paying Muse subscribers in the United States, it reached that stage after a September test limited to users who asked for access, and it arrives alongside a free-tier email feature that has not yet rolled out. The next signals to watch are how quickly the paid tiers convert, how businesses respond to the calls, and whether Meta publishes usage figures for the feature itself.
