Microsoft will begin rolling out a new in-meeting control for Microsoft Teams in July 2026 that lets licensed meeting organizers and presenters turn off all Meeting AI features during a session, according to the company's latest roadmap update. The move hands hosts a long-requested ability to disable tools like Copilot, the intelligent meeting recap, language facilitation, and other AI-driven experiences with a single toggle—even after a meeting has started.

What actually changed

The new control appears as a switch in a meeting's More actions menu, accessible only to users assigned the organizer or presenter role. When toggled off, it disables a bundle of AI-assisted features that normally run in the background: Copilot stops generating meeting notes, summarizing conversations, or answering prompts; Intelligent recap no longer creates automatic chapter divisions or speaker timelines; and real‑time aids like the Facilitator and Intelligent speaker view stop processing data. Once the toggle is flipped back on, the AI resumes working immediately—no need to rejoin or restart the meeting.

Microsoft has confirmed that the switch covers:

  • Copilot in Teams meetings
  • Intelligent meeting recap
  • Meeting Facilitator
  • Other intelligent meeting experiences that rely on live transcription or speaker recognition

The feature requires a Teams Premium or Copilot‑eligible license for the organizer, and it is specific to scheduled meetings; it does not appear in channel meetings or ad‑hoc calls. Because the control lives inside the meeting itself, it can be toggled mid‑session—a critical capability for discussions that start out casual but later touch on sensitive topics.

What it means for you

The new toggle reshapes how different roles experience Teams meetings. Here’s how it plays out for each group.

For everyday users and information workers

If you’re a meeting participant, you might suddenly lose access to an AI‑generated summary or the ability to query Copilot about what you missed. The organizer now has the final say on whether any AI processing happens. For people who have grown to rely on the recap or meeting notes to catch up after a conflict, that’s a significant shift. It makes transparent communication more important: organizers will likely announce at the start of a meeting whether AI is enabled, much as they already mention recording.

For presenters and organizers

Organizers gain a straightforward compliance tool. Sensitive discussions—salary negotiations, confidential project reviews, or client meetings under NDA—can run without worrying that transcribed content will linger in a meet‑ing recap or feed a large language model. Flipping the switch off also stops the AI from using any session data for real‑time processing. Importantly, the control is not available to “attendees” in a meeting; only those with elevated roles can flip it. That prevents a single participant from quietly disabling features others might expect to use.

For IT administrators and compliance officers

The toggle provides a user‑facing complement to the admin‑side policies already available in the Teams admin center. Admins can still set organization‑wide defaults for AI features or restrict them per user, but this per‑meeting control gives event organizers a pressure‑release valve when a specific session demands extra confidentiality. It also reduces the number of helpdesk tickets from employees who want a guarantee that the AI isn’t listening—they can now enforce that themselves during the meeting.

For developers and integrators

There’s no direct API change announced for this toggle, but any third‑party apps that depend on the Teams meeting transcription feed will be affected. When the organizer disables AI, the underlying transcription stream that many apps rely on for real‑time insights may stop. Developers building collaborative tools on the Teams platform should plan to handle a scenario where meeting transcription data suddenly becomes unavailable mid‑session, and they may want to surface a notice to users when that happens.

How we got here

Microsoft’s rapid injection of AI into Teams has been both a productivity boon and a privacy headache. The company introduced real‑time transcription in 2021, then layered on intelligent recap and Copilot in 2023 and 2024. While many organizations embraced these features, a vocal minority of users—particularly in regulated industries—raised concerns about data residency, attorney‑client privilege, and the simple discomfort of having every word captured and processed.

Past incidents have added fuel to that fire. In 2024, researchers found that over‑broad transcription settings could expose meeting content beyond the intended audience, and some enterprises reported that employees were hesitant to speak freely when an AI recorder was active. Microsoft responded with per‑user controls and tenant‑level policies, but until now there hasn’t been a single, easily discoverable button an organizer could press to shut everything off mid‑meeting.

The new toggle is the culmination of that feedback. It appeared on the Microsoft 365 Roadmap in late 2025 as feature ID 467250, and has since moved through targeted release rings ahead of the July 2026 general availability date. The messaging from Microsoft frames it as both a trust‑building step and a compliance enabler, acknowledging that AI in meetings isn’t always appropriate.

What to do now

If your organization uses Teams, a few practical steps will prepare you for the July 2026 rollout.

  • Check your licensing: The toggle requires the organizer to have a Teams Premium license or a Microsoft 365 Copilot license. Audit which users have those licenses now, and decide whether to expand licensing so that every potential meeting organizer can use the control.
  • Educate meeting organizers: Send a short guide to all employees who regularly schedule meetings. Explain where the toggle lives (More actions → Meeting AI), what it does, and, crucially, when to use it. Emphasize that turning off AI also disables helpful features like the recap, so they shouldn’t flip it lightly.
  • Update internal policies: If your company has a policy around recording meetings, extend it to cover AI features. For meetings that involve sensitive topics, consider making it a mandatory step for organizers to disable Meeting AI at the start.
  • Review admin defaults: In the Teams admin center, under Meetings → Meeting policies, you can set the default state for Copilot and intelligent recap. Decide whether you want those features on by default (so organizers actively have to turn them off) or off by default (and organizers turn them on only when needed).
  • Test before rollout: When the toggle reaches your tenant’s preview ring, run a few test meetings. Verify that the AI stops and resumes cleanly, and that any third‑party integrations behave as expected. File feedback with Microsoft if you encounter glitches.

Outlook

Microsoft is expected to deliver the toggle to worldwide commercial tenants by late July 2026, with GCC and GCC High clouds following in August. The feature marks a maturing of the company’s AI governance story—giving control back to users after an era of aggressive feature push. Watch for similar controls to propagate to other Microsoft 365 apps; Outlook and Word have already tested versions of a “disable Copilot” button, and meetings were the logical next step. Longer term, the ability to turn AI on and off at a granular level could become a negotiating point in enterprise contracts, as compliance departments demand more assurances about where and when machine learning touches corporate data.