Meta is giving Muse a voice and a visual presence. Explore the official announcement and the questions Latin American businesses should ask before planning new customer conversations with AI.
Daniella Jovel Head of Performance & Media Strategy
Economist by profession, focused on maximizing digital profitability and business ROI.
Meta AI Video Calls: What Muse Means for Brands
What changes when an AI assistant has a visible presence on the other side of a call? For a business, the question opens a discussion about designing experiences that feel approachable and help someone make progress. A voice and a character may make an introduction easier. The quality of the help will determine whether a customer wants to return.
On November 5, 2026, Meta announced voice and expressive avatars for Muse, starting in the United States across mobile, web and Mac. Meta AI video calls bring a visual presence to conversations with its personal agent.
Official introduction to Muse. Source: Meta.
What Meta announced and what the experience involves
The official announcement describes voice-only conversations, an avatar beside the chat, or a full-screen view, with the ability to interrupt an answer. The rollout starts in the US.
The participant on the other end is an AI agent. That sets a different expectation from a meeting with a colleague or a service representative. We can assess whether it explains clearly, follows instructions and helps complete a task. A persuasive appearance should not become the reason we assume it has understood every detail.
At Meta Connect 2026, the company also outlined plans to bring Muse to its AI glasses and showed Muse Charm. Those plans provide context for a broader product direction; actual access needs to be checked when considering a specific use.
For marketing teams, that distinction helps keep planning grounded. A product demonstration can suggest an interesting customer experience, while a campaign still needs to work with what its intended audience can use. Start by identifying the communication problem a spoken, visual interaction might help solve.
How an avatar gives a conversation a sense of presence
Meta AI Research describes Muse Realtime Voice and Muse Realtime Avatar as a shared streaming system. Speech information helps coordinate the voice, lip movements and expressions, while reference media establishes the character’s appearance for the generated visual performance.
Our editorial interpretation is that this coordination could make an explanation easier to follow. A pause or an apparent listening gesture adds a cue to the exchange. Presentation quality and answer quality still need separate assessment: a character can appear attentive while missing an important condition in the question.
Consider a usability exercise for a service explanation. After the conversation, can the participant describe what the service includes and what to do next? Do they understand that they interacted with AI? If the animation is convincing but the customer remains confused about the offer, the content and the next step need attention.
What this could mean for a brand
Meta’s introduction to Muse describes a personal agent that can carry out tasks and work with applications according to the access a person grants. That is useful context for understanding why the company is adding a voice and a visible character.
For a brand, we would use the announcement to explore a few practical questions. Where do customers struggle to understand your offer? Which explanations require several follow-up messages? At what point does someone need a person who can consider their individual situation? These are strategy exercises rather than business features confirmed by this release.
Making an offer easier to understand: rehearse answers about scope, deliverables and the steps required to become a customer. Aim for an informed decision at the end.
Helping a team prepare: practise a sales introduction or a difficult conversation using fictional scenarios, then review the parts that need a clearer explanation.
Improving the support journey: identify questions that suit a guided answer and situations that require human judgment, additional context or follow-up.
This work can be useful before a team adopts any avatar. When a business struggles to explain what it sells, a visual character will carry that same uncertainty into the conversation. Organising the information and choosing a useful next action are practical places to begin.
What teams in Latin America should watch
For a Latin American business, the US rollout is a reference point for learning. Before planning a local trial, check access to the specific feature for the account and country involved. Access to a wider product should not be treated as proof that every feature is available under the same conditions.
A useful evaluation would include the language customers actually use: regional expressions, local product names and common ways of asking for help. It should also examine what happens when someone changes their mind, speaks with background noise or asks for an explanation to be repeated.
Compare a future trial with the channel customers already use. A short written answer may resolve some questions with less effort. Other situations may benefit from a spoken exchange. The choice should follow the task and the preferences of the person seeking help, with a clear way to continue when the first answer falls short.
How to prepare a useful first experiment
We recommend starting with a defined business question and a small content exercise. You can prepare the information and evaluation criteria while you check whether the relevant feature is available for your audience.
Pick one customer conversation: choose a question your team regularly receives and define what the customer should understand when the exchange ends.
Prepare an approved answer: include the relevant exceptions and the facts that must be confirmed before offering guidance.
Make the identity clear: for any AI experience you test, explain who is providing the answer and how someone can reach a person.
Use fictional test cases: try incomplete questions, changes of intent and situations that require stopping to ask for more information.
Evaluate progress: look at understanding, resolution and repeat questions. Record what improves compared with the experience you already provide.
Meta AI video calls raise a useful design question: what should a visible assistant help a customer accomplish? Exploring that question calls for clear expectations, a consistent voice and well-prepared information. A good first experiment begins with the customer’s need and gives the team evidence for deciding what to try next.
Official sources
Product details are supported by the following official Meta publications, accessed on November 9, 2026. Business scenarios and recommendations are La Fábrica’s editorial analysis.