Microsoft Teams Introduces Facilitator, an AI That Listens to Meetings

Microsoft is preparing a new feature for Teams called Facilitator, designed to continuously listen to the course of a meeting and fill gaps in participants' knowledge on its own. When someone asks a question that goes unanswered or expresses uncertainty, the assistant will search the internet itself and paste the result into the meeting chat, before anyone even asks.

How Facilitator Works

According to Microsoft's description, Facilitator recognizes the moment when participants ask about something nobody answers, or signal a lack of knowledge on a topic. The system then analyzes the conversation's context and the meeting agenda, searches the web, and pastes a ready-made answer directly into the meeting chat. The assistant doesn't speak up or interrupt the conversation, all of its activity is limited to text in the chat window.

Facilitator can recognize when participants are asking questions or expressing uncertainty, and then searches for and shares relevant answers using web search in the meeting chat - from Microsoft's announcement

The feature requires the organization to have Copilot's web search enabled. If an administrator turns that option off, Facilitator stops working regardless of the individual participant's settings. That means the decision to activate the assistant rests, in practice, with IT departments rather than individual employees.

Continuous Listening and Privacy Questions

The mechanism itself raises the most concerns: for Facilitator to be able to respond at any point during a meeting, it must continuously process the audio and content of the conversation. Microsoft doesn't hide that this is a necessary condition for the feature to work, but the company has so far not clarified two key issues for security and compliance departments.

The first concerns whether data collected during meetings with Facilitator enabled will be used for further training of Microsoft's models. The second is the lack of warnings about possible hallucinations, meaning situations where the system pastes incorrect or made-up information into the chat, presenting it as fact. For regulated companies, whether legal or medical, this isn't a cosmetic issue, since an incorrect answer will end up in the meeting record on equal footing with points made by actual people.

Response to Criticism

After the first reports about the feature, Microsoft modified its approach to controlling AI assistants in Teams. The company announced that users will be able to turn off Copilot, Facilitator, and automatic Recap notes during an ongoing meeting, not just before it starts. This is a response to criticism from users who pointed out that current AI settings in Teams are hard to change on the fly once a conversation is already underway.

What This Means for Companies in Poland

For Polish organizations using Teams and the Microsoft 365 Copilot Premium package, what will matter most is who within the company decides to enable Facilitator and under what rules. Since the feature requires administrator consent at the tenant level, it will be IT and security departments, not individual employees, who have to assess the risk tied to continuously listening in on business conversations, especially in the context of RODO (Poland's GDPR implementation) and the confidentiality of client data.

The rollout of such features coincides with growing regulatory pressure for transparency in AI systems across the European Union. Companies that decide to enable Facilitator will need to show the legal basis on which they process meeting recordings and transcripts, a question that so far remains without a clear answer from Microsoft.

Expanding tests to Targeted Release starting in early August 2026 will make it possible to assess how the feature performs in practice before it reaches all Copilot Premium customers by the end of the month. Until then, administrators can prepare their own usage rules before the feature appears as an option to enable in the Teams admin console.