Meta Muse AI Agent Adds Video Chat and Avatars for US Users

What happened

Meta has made its Muse AI agent available to US users for video chat, letting them talk with the assistant through a customisable animated avatar. The feature can also take calls, according to TechCrunch's reporting, and the rollout arrives ahead of a planned Tamagotchi-like wearable designed to bring Muse to life.

The update moves Muse from a text-and-voice assistant towards a visual, persona-led interface. Users can shape how the avatar looks, which makes the agent a recognisable character rather than a generic chat window. The source material does not set out pricing, eligibility limits or a wider international timeline, so those details should be treated as unconfirmed.

Why it matters

For organisations building AI into customer-facing services, the development shows how quickly conversational AI is shifting from a utility into a presence. A face, a voice and a personality change what people expect: they tend to judge an agent by its tone and responsiveness as much as by the accuracy of its answers. Any team deploying AI in service channels now has to design for that expectation, not just for task completion.

The planned wearable adds a second dimension. A physical companion that depends on a persistent AI agent raises the question of how always-on interaction should be designed, governed and switched off. Those are service-design questions as much as product questions, and they will carry over into any organisation that puts AI into devices people carry or wear.

The Renascence take

Most readers will focus on the novelty of an avatar that takes calls. The more useful question is whether people feel a relationship with the agent, and whether that feeling is earned by good outcomes or manufactured by good animation.

An avatar makes an AI agent easier to like, but it does not make it more useful, and likeability can quietly lower a customer's guard. The behavioural principle underneath is trust calibration: people extend trust to faces and personalities faster than their experience justifies. A customer-obsessed operator should therefore test avatar-led interactions against resolution rates, escalation paths and clear disclosure that the customer is talking to AI, before copying the format for its own channels. Treat the character as the last layer of design, not the first.