CONFIRMED
Meta announced on November 5 that it is rolling out voice conversations and expressive avatars in Muse, its personal artificial intelligence (AI) agent. The change lets users discuss an idea and receive spoken responses without continually typing instructions. It is coming to the mobile app, the web and Mac, but the rollout starts in the United States. The announcement does not confirm availability in Brazil.
From a text box to a conversation
Meta instructs users to tap the phone icon in the upper right corner of Muse to begin. The company says conversations support interruptions and offer hundreds of voices with different styles. Users can choose audio alone, keep an avatar alongside the conversation or switch to a full-screen view. The character reacts while listening and responds with expressions and gestures tied to the tone of the exchange.
A rollout, not a brand-new research reveal
Meta previewed this experience at Connect 2026. In September, its research team detailed Muse Realtime Avatar, which generates a visual performance from the same stream used for speech. Independent coverage by AI Insiders explained that architecture and warned that a technical demonstration does not necessarily mean a feature is available inside the app. Thursday's material change is the explicit announcement of a product rollout, rather than the discovery of an entirely new underlying model.
A rehearsal before the meeting
For a small business, the clearest potential benefit is not an animated face on a screen. It is the ability to rehearse a sales presentation, explore a customer's objections or organize an idea by speaking. Meta itself cites interview preparation, presentations and trip planning. A professional could practice explaining a service and request difficult follow-up questions before talking to a real client. These are possible applications, not productivity results measured by MaxAssistant. The useful comparison is with the time and effort required to reach the same outcome through text.
A fluid exchange is not guaranteed execution
Meta also says Muse can work in the background while a user talks. That does not prove every assignment will be completed correctly or make the app a ready-made customer service system for a Brazilian business. The announcement does not present an interface for integrating these avatars into third-party customer support, establish savings per task or confirm local availability. Before sharing customer information, a business needs to check data policies, permissions and applicable terms. The technical research also states that Muse is intended for users aged 18 and older.
The test is operational, not visual
The development matters because it can reduce the friction between thinking and expressing a request. A convincing facial expression, however, does not itself improve factual accuracy. Our view is that a useful evaluation should measure understanding, required corrections and time to a usable deliverable. For Brazil, this announcement signals product direction rather than immediate adoption. This report was initially published as a rumor on October 25 and has been updated with the official confirmation of voice and expressive avatars. Avatar sharing is a separate subject and is not confirmed by this announcement.
Sources
Meta: Give Your Muse a Voice, November 5; Meta AI Research: Bringing Your Muse to Life, September 23; AI Insiders: technical context and limitations of Muse Realtime Avatar, September 25.
