Meta’s Consumer AI Agent Muse Gains the Ability to Make Phone Calls for U.S. Subscribers

After initially testing the feature in September, Meta’s consumer AI agent Muse is now officially able to place phone calls to U.S. businesses on behalf of its users. However, the company has clarified that the capability is currently restricted to paid subscribers, marking another significant step forward in Meta’s ongoing effort to integrate its artificial intelligence offerings deeper into the daily lives of everyday consumers.

Meta’s Muse, which the company announced last month has surpassed 10 million users, has been positioning itself as a mainstream product designed to handle routine digital and logistical tasks. Placing outbound phone calls represents one of the most ambitious ways the platform is attempting to achieve this goal, as it directly targets some of the most friction-heavy and tedious items on a typical person’s to-do list.

Alexandr Wang, Meta’s chief AI officer and head of the company’s Superintelligence Labs division, highlighted the practical applications of the new capability on social media. According to Wang, the feature can be leveraged by users to take care of phone calls they would otherwise prefer to avoid, such as negotiating with a cable company, navigating the frustrating automated systems of the DMV, or booking a table at a restaurant that does not offer online reservation options.

Previously, the capability was restricted to a select cohort of Muse early adopters who had to actively request access to the experimental feature. With the broader rollout, any consumer paying for a Muse subscription can initiate these requests. Currently, subscription tiers for Meta’s AI agent start at $20 per month, which grants users an allocation of 500 million Muse tokens per week. For power users and heavier consumers of artificial intelligence capabilities, Meta also offers a $100-per-month subscription tier that provides up to 3 billion tokens.

Muse is far from the only player in the burgeoning space of AI-driven phone assistants. Rival startups such as Instinct have rolled out similar calling features, and tech giants like Google have begun testing comparable functionalities, such as allowing Gemini to make phone calls initially for U.S. Pixel owners.

Despite the rapid technological progress enabling AI models to carry out these conversations, receptionists and businesses on the receiving end of these automated calls have responded with mixed reactions. Reports indicate that many businesses and customer service representatives hang up on AI agents almost immediately upon realizing they are speaking to a synthetic voice rather than a human customer. Nevertheless, as artificial intelligence models continue to improve in their ability to mimic natural human inflection, pacing, and conversational nuances, distinguishing between a flesh-and-blood customer and an AI proxy is becoming an increasingly difficult task for human operators.

These hurdles notwithstanding, the feature holds profound appeal for a broad consumer base that has grown increasingly exasperated by the steady decline of traditional customer service standards over the years. For decades, consumers have been forced to endure complex automated phone trees, navigate endless mazes of routing options, and sit on hold for extended periods waiting for a human customer service representative to answer. Even when a representative finally picks up, callers frequently face the risk of being transferred across multiple departments or finding that the person on the other end lacks the authority or capability to resolve the issue. Now, consumers have the option to deploy an AI agent with virtually infinite patience to shoulder these tedious burdens on their behalf.

While voice calling capabilities remain strictly behind a paywall for the time being, Meta is also expanding its ecosystem in other directions. Alexandr Wang noted this week that a recently introduced feature allowing users to create a dedicated Muse email address will soon be rolled out to free users as well, signaling Meta’s balanced approach to tiering its consumer-facing features.