New Controls for Copilot in Teams: Helpful Clarity or Growing Complexity?

Do you ever want Copilot to step out of your meeting for a few minutes so you can discuss what you need without it listening? With a new Copilot disabling option in Teams, you'll be able to do this soon (the new feature is rolling out by October).

On one level, more control like this is great. But on another, combined with another update I recently highlighted - the default Copilot mode changing to Copilot without transcript - understanding what's going on with recording, transcription, and Copilot in meetings is potentially getting more confusing.

As I understand this update (MC1139485), the Copilot "Turn off for everyone" option will disable all of Microsoft's inbuilt AI experiences and the associated collection of assets needed to facilitate them. Copilot will be disabled, Facilitator will be turned off, and if recording or transcription are running, they will be disabled.

However, as previously reported, another update, MC1139493, will change the default options for new meetings so that Copilot activation enables "Copilot without transcript". In this case, Copilot is active solely within the bounds of the meeting, relying on an ephemeral transcript that isn't preserved after the meeting.

How do these two changes likely interact with one another?

Irrespective of which mode you are using Copilot in, either without a transcript, or with a transcript (either because you changed your meeting options, or your default has been changed in a Teams meeting policy, or because you manually turned on recording/transcription, or because you enabled a feature like Facilitator that requires transcription to be turned on), the "Turn off for everyone" option will totally disable all those experiences until one or all is re-enabled in that meeting.

To my current understanding, if you are in a meeting that has Copilot enabled while transcription is turned off, the only mechanism available to pause or disable Copilot's knowledge of the meeting from the first point it was activated will be this option. However, if your meeting requires transcription to use Copilot, simply turning off transcription/recording has the same effect.

For most organizations where these different scenarios become relevant are edge cases. It's not every day that this level of minutiae about how Teams or Copilot is handling your content is relevant to your work. But I can imagine scenarios or highly regulated organizations where understanding this complexity is relevant.

As a foundation, highlighting these features to your users, so they can fully understand where their words in meetings, or those of them people they are meeting with, might be available to others is vital. This principle of transparency is essential to responsible AI use.

Is understanding what Copilot in Teams is doing or can see too confusing? Do we need a paradigm more like other AI notetakers, where they are either in the meeting as a user or booted out?

Posted on Linkedin on 08/28/2025 -> View Linkedin post here