Meta has expanded the capabilities of its personal AI agent, Muse, allowing paying subscribers in the United States to make phone calls to businesses on their behalf. The feature can handle everyday tasks such as contacting insurance companies, speaking with cable providers, checking appointment availability and making restaurant reservations. The announcement was made on November 12, 2026, by Meta Chief AI Officer Alexandr Wang.
The rollout marks another step in Meta’s effort to turn Muse into a personal digital assistant capable of completing tasks beyond answering questions and generating content. Previously, phone calling was available only to a limited group of early users who had requested access. The feature is now being made available to paid subscribers in the US, with subscriptions starting at $20 per month. However, the service still faces challenges, including businesses that refuse to interact with AI callers and questions about how AI-generated phone conversations should be disclosed.
Meta Muse Can Now Make Phone Calls for Users
Meta’s Muse can now place outbound calls to US businesses and communicate with employees or customer-service representatives to complete tasks requested by users.
The feature is designed to handle phone calls that people often find inconvenient or time-consuming.
For example, users can ask Muse to:
- Contact an insurance company about a policy.
- Call a cable or internet provider regarding billing issues.
- Check appointment availability at a business.
- Contact restaurants that do not offer online reservations.
- Ask businesses about products and services.
- Make enquiries that would otherwise require waiting on hold.
The AI agent can conduct the conversation and attempt to complete the requested task without requiring the user to make the call personally.
Meta Chief AI Officer Alexandr Wang announced the wider rollout on November 12, describing it as a feature users had frequently requested.
The company had initially tested phone calling with selected early adopters in September. That limited rollout allowed users who specifically requested the capability to try it before broader access was introduced.
Muse Phone Calling Is Available Only Through Paid Plans
Unlike some of Muse’s basic capabilities, phone calling is currently restricted to paying subscribers in the United States.
Meta’s subscription plans offer different weekly token allowances depending on how heavily users rely on the AI agent.
| Subscription plan | Monthly price | Weekly token allowance |
|---|---|---|
| Power | $20 | 500 million tokens |
| Maximum | $100 | 3 billion tokens |
The $20 plan is the entry-level paid subscription, while the $100 option is designed for users who require significantly greater usage capacity.
The token allowance determines how much AI usage subscribers can access across the service. Meta has not publicly specified how many tokens an individual phone call consumes or how many calls are included in each subscription.
This means the actual cost of using Muse for frequent business calls will depend on overall usage and the resources consumed by other AI tasks.
For now, users who rely on the free version cannot access the phone-calling feature.
How Does Muse Handle Business Calls?
The calling capability allows Muse to act as an intermediary between a user and a business.
A user can describe the task they want completed, after which the AI agent attempts to contact the relevant business and communicate the request.
For example, someone looking for a restaurant reservation could ask Muse to call a restaurant, check available tables and communicate the preferred booking details.
Similarly, a user dealing with a complicated cable bill could ask the agent to contact the provider and enquire about charges or available options.
The feature could be particularly useful for situations involving long waiting times, complicated customer-service menus or businesses that do not support online booking.
An earlier demonstration shared by Meta Superintelligence Labs employee Matt Schlicht showed Muse calling a Volvo dealership to enquire about appointment availability. The agent reportedly waited on hold for approximately 15 minutes before obtaining information.
Such examples illustrate how AI agents could take over routine communication tasks that traditionally require human attention.
However, the ability to make a call does not guarantee that every request will be completed successfully. Businesses may refuse to speak with AI agents, and complicated negotiations or sensitive requests may still require human involvement.
Meta Muse Has More Than 10 Million Users
The phone-calling rollout comes as Meta continues expanding Muse’s consumer user base.
According to the company, Muse had attracted more than 10 million users by October 2026.
The growth highlights Meta’s ambition to establish a personal AI assistant that can manage different aspects of everyday life.
Unlike traditional chatbots that primarily respond to questions, Muse is designed to interact with external services and complete tasks with user permission.
Its capabilities include:
- Connecting with email and calendar services.
- Managing everyday administrative tasks.
- Supporting shopping and travel-related activities.
- Helping users organise information.
- Interacting with connected applications.
- Making purchases through supported services.
- Contacting businesses through phone calls.
Meta is also exploring ways to connect Muse more closely with its wider ecosystem of consumer and business services.
The company’s strategy is to make AI assistance useful for practical activities rather than limiting it to conversational interactions.
Businesses Are Not Always Willing to Speak With AI Agents
Despite the convenience, AI-powered phone calls face an important challenge: businesses may not want to interact with automated callers.
Some receptionists and customer-service representatives reportedly end calls when they realise they are speaking with an AI agent rather than a human customer.
This can prevent the agent from completing even relatively straightforward tasks.
The challenge is particularly relevant for businesses that have concerns about automated calls, identity verification, fraud or the handling of sensitive customer information.
Meta’s calling feature also raises questions about disclosure. Businesses may want to know whether the person on the phone is a human customer or an AI acting on someone’s behalf.
A successful AI calling system must therefore do more than produce natural-sounding speech. It needs to communicate accurately, respect business policies and handle personal information responsibly.
Meta Previously Tested a Human-Assisted Calling System
Muse’s calling feature has also attracted attention because Meta previously experimented with a human-assisted approach.
In September 2026, Reuters reported that Meta had tested a human concierge system in which contractors could place certain phone calls on behalf of Muse users.
The experiment was intended to address situations where businesses repeatedly refused to communicate with AI agents.
According to the report, internal tests suggested that human-assisted calls could achieve success rates of approximately 95% to 98%, higher than the reported success rate of fully automated calls.
However, the experiment raised privacy and transparency concerns within Meta.
Employees questioned whether sensitive information could be exposed to contractors and whether users were adequately informed that a human might handle their calls.
A Meta executive acknowledged shortcomings in the disclosure process, and the company said the feature had been rolled back while it worked on appropriate safeguards.
The distinction is important: the latest public rollout is the AI calling capability, and the earlier human-concierge experiment should not be confused with the standard automated calling feature.
Meta Faces Competition From Google and AI Startups
Meta is not the only technology company exploring AI-powered phone calls.
Google has introduced similar capabilities through its AI products, while startups are also developing agents that can contact businesses and complete tasks.
One example is Instinct, which introduced its Concierge feature to help users make restaurant reservations, contact service providers and handle other high-touch requests.
These developments reflect growing competition in the personal AI agent market.
Companies are increasingly trying to build assistants that can perform tasks across websites, applications, email services and telephone networks.
The competition is moving beyond the quality of AI-generated responses towards the ability to complete real-world activities reliably.
For Meta, its large consumer ecosystem could provide an advantage in attracting users and connecting AI agents with businesses.
However, widespread adoption will depend on whether businesses are willing to accept AI-generated interactions.
Meta’s Bigger Ambition: Turning AI Into a Personal Assistant
The calling feature is part of Meta’s broader strategy to develop Muse into a digital assistant capable of managing everyday activities.
The company wants users to rely on AI not only for information but also for practical tasks that normally require several steps.
This could eventually include contacting service providers, organising appointments, communicating with businesses and managing routine administrative work.
Meta is also exploring ways to monetise the platform through subscriptions, business services and commerce-related transactions.
Paid subscriptions provide a direct revenue opportunity, while business integrations could create additional commercial possibilities.
The wider opportunity is significant because phone calls remain an important part of customer service despite the growth of online booking systems and messaging applications.
An AI agent capable of handling these conversations could save users time, particularly when dealing with businesses that still rely heavily on telephone support.
The Bigger Picture
Meta’s decision to make Muse’s phone-calling capability available to paid subscribers represents another step in the development of agentic AI. Instead of simply providing information, AI systems are increasingly being designed to interact with businesses, communicate with people and complete practical tasks. For consumers, this could reduce the time spent waiting on customer-service lines, making appointments and handling routine enquiries.
However, the technology still faces important limitations. Businesses may reject automated callers, sensitive conversations require appropriate safeguards, and users need transparency about how their requests are handled. Meta’s earlier human-concierge experiment also demonstrated the privacy and disclosure challenges involved in allowing AI services to operate through telephone networks. The long-term success of Muse will depend on whether it can make these interactions reliable, secure and acceptable to businesses as well as consumers.
Looking Ahead
Meta is expected to continue expanding Muse’s capabilities as it competes with Google and specialised AI agent startups. The company’s immediate challenge will be improving the reliability of automated calls and ensuring that businesses can identify and appropriately handle AI-generated interactions. Subscription pricing, token consumption and the practical success rate of calls will also influence whether users consider the feature worth paying for.
Over the longer term, AI agents could become an additional channel through which consumers interact with businesses. Rather than navigating phone menus or waiting for customer-service representatives, users may increasingly delegate routine conversations to digital assistants. For that model to become mainstream, technology companies and businesses will need to establish clear standards for consent, identity verification, privacy and communication between AI agents and human representatives.
SEO Metadata
SEO Title: Meta Muse Can Now Call US Businesses, but Only Paid Users Can Access It
Meta Description: Meta expands Muse’s AI phone-calling feature to US paid subscribers, allowing the assistant to contact restaurants, insurers and service providers.
URL Slug: meta-muse-ai-agent-calls-us-businesses-paid-subscribers
Focus Keyphrase: Meta Muse phone calling
Secondary Keywords: Meta Muse AI agent, Muse paid subscription, Meta AI phone calls, Alexandr Wang, AI business calls, AI personal assistant, AI agent calling feature, Meta Muse pricing
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
