Meta’s Muse AI Can Now Phone US Businesses for Paid Subscribers

November 15, 2026

Meta's Muse AI Can Now Phone US Businesses for Paid Subscribers

Meta Muse AI phone calls are now a reality for paying subscribers in the United States. Meta’s consumer AI assistant, Muse, can place outbound calls to American businesses on a user’s behalf, taking over the everyday phone errands most people put off for days: booking a table, chasing an insurance claim, sorting out a cable bill, or reaching the motor vehicle department. The feature was announced on 12 November 2026 by Meta’s chief AI officer, Alexandr Wang, and it marks a clear step in the race to turn chatbots into assistants that act in the real world, not just answer questions.

Muse is Meta’s personal AI agent, built to carry out tasks from a user’s instructions. With outbound calling, a user describes the errand in a chat, and Muse rings the business and reports back.

Meta Muse AI phone calls: what changed on 12 November

Meta had been testing business calls since September with a small group of early users who had to ask for access. What changed on 12 November is the audience: the feature is now open to any paying Muse subscriber in the United States, according to TechCrunch. Free users are still locked out, and so is everyone outside the US. Meta has not said when calling will reach free users or subscribers in other countries.

The jobs Muse is pitched for are deliberately ordinary. Wang’s announcement listed the calls people least want to make: the insurer, the cable company, the motor vehicle department, and the restaurant that still takes bookings by phone. These calls are not difficult; they are time-consuming, often involve waiting on hold, and are easy to postpone — exactly the kind of chore an assistant is supposed to absorb.

How much it costs

Outbound calling sits behind Muse’s paywall. Plans start at $20 a month for 500 million Muse tokens a week, with a heavier tier at $100 a month for 3 billion tokens, according to TechCrunch. Meta has not published a separate price per call, nor said whether heavy callers face a separate limit, so check the terms before subscribing mainly for calling.

The pricing shows how Meta plans to make money from Muse, which now has more than 10 million users, according to the company. A feature that saves real time is a stronger reason to pay than a faster chatbot.

The limits of an AI on the phone

The feature now on general release is, as far as public reporting goes, fully automated. But its usefulness depends on something outside Meta’s control: whether the business on the other end stays on the line. Some businesses end calls once they realise they are talking to an automated system, according to TechCrunch — a real obstacle that no amount of clever software can fix on its own.

There are also questions about how much Muse can complete in a single call. A restaurant reservation is a simple exchange of a name, a time and a party size, but an insurance query can involve account numbers and authorisation a customer may not want to hand to an assistant. Meta has not documented every limitation, so the practical boundaries will only emerge as more people use it.

Privacy is part of that picture. Meta had earlier tested a version of the feature in which human contractors stepped in to help with calls Muse struggled to complete, and the company pulled that arrangement back after privacy concerns among employees about users unknowingly having conversations overheard by people rather than software. That history helps explain why the company now presents the feature as fully automated.

The wider race to build agents that act

Muse is not the only AI assistant learning to make calls. The startup Instinct already offers a similar option, and Google’s Gemini can also place calls to businesses. What is new is the speed at which the biggest AI companies are moving in the same direction: away from chatbots that wait for a question, and towards agents that keep working on a task without being asked again. OpenAI has been rolling out Dots, its always-on agents that run on their own cloud computer and connect to thousands of apps, and Anthropic launched Claude Meetings, which adds an AI participant to workplace video calls to take notes and track action items.

What it means for ordinary users

For now, the practical effect is narrow. The feature is limited to paid subscribers in the United States, and it only calls American businesses. Most readers outside the US, including in Pakistan, cannot use it yet, and Meta has given no wider rollout date. But the direction of travel is clear.

First, the value of an AI assistant is shifting from what it knows to what it can do. Answering a question is useful; completing a chore is worth paying for. If Muse’s calling works reliably, it will set a standard every other assistant is measured against. The same shift is visible in everyday money: in Pakistan, digital payments have grown so fast that the State Bank has been asked to prepare a merchant discount rate scheme for Raast QR payments.

Second, trust. Handing an AI assistant the phone means handing it your name, your plans and sometimes your account details. Users will need clear rules about what the assistant may say and what it must ask permission for first.

What this means

Meta’s Muse can now make real phone calls to US businesses for paying subscribers, starting at $20 a month. It is a small feature with a large signal: AI assistants are moving from answering questions to doing errands, and the companies behind them are betting that saving time is worth paying for. The limits are real — some businesses hang up on automated callers, and the feature is US-only for now — but the race it belongs to is global. Bankopedia will keep covering these developments.

Sources