Meta Rolls Out AI Phone Calling Feature for Paid Muse Subscribers
Meta has launched an outbound phone calling feature for paid subscribers of its Muse AI agent in the U.S., enabling automated handling of calls to businesses and service providers.
Meta has officially rolled out a feature for its consumer AI agent, Muse, enabling it to place phone calls to U.S. businesses on behalf of users. Initially tested in September, the function is now available across the United States, though Meta has restricted access to paying subscribers. The company, which revealed last month that Muse has accumulated over 10 million users, is positioning automated calling as a practical feature to help make consumer AI a part of everyday life.
Alexandr Wang, Meta’s Chief AI Officer and head of its Superintelligence Labs division, highlighted the capability as a solution for managing tedious phone errands. In a public announcement, Wang noted that the tool is designed to handle phone calls that consumers typically avoid, such as contacting cable companies, insurance providers, the Department of Motor Vehicles (DMV), or booking dinner reservations at restaurants that lack online booking systems.
Tiered access and market context
During the initial testing phase, the calling feature was available only to a select group of early adopters who requested access directly through the AI. Under the updated rollout, any user with a paid Muse subscription can utilize the service. Meta currently offers two main subscription tiers: a base plan at $20 per month providing 500 million Muse tokens per week, and a higher tier priced at $100 per month offering 3 billion tokens per week for heavy users. In addition to calling, Wang recently noted that a feature allowing users to create a Muse email address will soon become available to free tier users.
Meta is not the first company to explore AI-driven outbound phone calls. Google’s Gemini and the startup Instinct already offer similar functionality. However, the adoption of outbound AI calling faces operational challenges on the receiving end. Many businesses and receptionists routinely disconnect phone calls once they realize they are speaking with an automated agent rather than a human caller. As AI voice synthesis technology improves and mimics human speech patterns more convincingly, distinguishing between an actual customer and an AI agent acting on their behalf is becoming increasingly complex.
What it means for developers
The expansion of outbound calling in consumer agents like Muse signals a growing shift from text-only chatbots toward active, agentic task completion. For developers building next-generation applications, this trend highlights the importance of integrating multi-modal voice capabilities, structured dialogue management, and resilient error-handling into autonomous agents.
As consumer platforms push token-heavy voice and calling interactions behind paid subscriptions, developers face similar challenges in managing infrastructure costs and model access. Building, testing, and benchmarking voice agents or autonomous workflows requires access to diverse language models. Developers looking to experiment with multi-model architectures can try top AI models cheaply through one API at https://apixoai.online, simplifying integration across different provider ecosystems.
Furthermore, developers building business-facing systems will need to account for AI-to-AI and AI-to-human call handling. As AI agents increasingly initiate contact with service providers, software engineers working on call center platforms or IVR (Interactive Voice Response) systems will need to adapt call screening protocols, verification methods, and data intake mechanisms to handle automated callers efficiently.
Addressing customer service friction
From the consumer perspective, automated calling addresses longstanding frustrations associated with customer service channels. For years, callers have had to navigate multi-layered phone menus and endure lengthy wait times before reaching a human representative. Outbound AI agents offer infinite patience, allowing users to delegate hold times and basic inquiries entirely to software.
Despite pushback from call recipients who prefer human interaction, the convenience of delegating routine administrative tasks remains a major selling point for consumer AI products. As Meta and its competitors refine these outbound capabilities, automated phone handling is likely to become a benchmark feature for premium AI subscription tiers.
Source: Muse can now place calls to U.S. businesses — if you pay — TechCrunch AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
