Microsoft is handing meeting organizers the switch they’ve been asking for: the ability to disable artificial intelligence during a live Teams call. Starting in July 2026, any licensed meeting organizer or presenter will be able to toggle off Copilot, the Facilitator agent, and Intelligent Recap with a single in-meeting control, without disrupting the call.

Dubbed “Meeting AI,” the control represents a quiet but pivotal shift in how Microsoft balances AI productivity with user autonomy—and it arrives after months of feedback from enterprises that demanded more granular, real-time governance over sensitive discussions.

How the new Meeting AI toggle works

The Meeting AI control will appear inside the meeting toolbar for organizers and presenters who hold the appropriate license. With one click, it overrides the default AI settings for that specific session. When turned off, three features go dark immediately:

  • Copilot: The AI assistant that answers questions, summarizes live discussions, and generates content based on meeting transcripts stops responding and stops capturing new data.
  • Facilitator: The agent that manages the meeting agenda, sends real-time prompts, and nudges participants loses its ability to monitor or intervene.
  • Intelligent Recap: The post-meeting feature that automatically generates meeting notes, highlights, and task assignments will no longer produce a recap for that session.

Importantly, toggling AI off during a meeting does not retroactively delete data already captured—transcripts or Copilot interactions that occurred before the toggle were stored according to the tenant’s existing policies. The change only prevents further collection and processing.

The toggle is live and bidirectional: an organizer can switch AI back on later in the same meeting if the conversation shifts and participants agree. Microsoft has not yet detailed whether other participants will see a notification when AI is disabled, but the company’s typical approach suggests a transparent indicator will appear in the meeting experience.

Who gets the control and when

The meeting AI toggle is scheduled for general availability in July 2026, according to Microsoft’s roadmap. It will require a license that includes Microsoft 365 Copilot or Teams Premium; standard E3 or E5 licenses alone will not unlock it. Organizers and presenters—not every attendee—will see the toggle when those licenses are present.

Microsoft’s decision to gate the control behind a license aligns with its broader strategy of tying advanced AI governance to its premium offerings. For organizations already paying for Copilot, the toggle arrives as a value-add; for those on standard plans, it may become another reason to upgrade.

The feature is expected to light up first in the new Teams client on Windows and Mac, with mobile parity following shortly after. Web-based Teams will likely get it later, as Microsoft often prioritizes desktop builds for meeting-experience features.

What this means for meeting organizers and presenters

For the person running the meeting, the toggle transforms them from AI bystander to AI gatekeeper. Consider a few scenarios:

  • Sensitive HR discussions: An organizer can ensure that no Copilot transcript or Facilitator prompts capture employee relations details.
  • Strategic planning sessions: Turning off Intelligent Recap prevents AI-generated summaries from being stored and later surfaced for unauthorized users.
  • Brainstorming without filters: Teams may want to ideate freely without an AI agent inserting suggestions or storing every word.

The control also relieves a social friction point. Earlier this year, many users expressed discomfort when Copilot silently joined meetings; some declined to speak candidly. Now, an organizer can explicitly switch AI off, signaling to participants that they’re in a walled-garden conversation—no agents listening, no recaps saved.

However, the toggle is not a fire-and-forget tool. Organizers must remember to use it. If someone forgets to disable AI at the start of a confidential meeting, Copilot will already have ingested the initial minutes. Microsoft has not indicated plans for a pre-meeting default-off setting, so the onus remains on the host.

For IT administrators: less dependence on sweeping policies

Until now, controlling AI in Teams meetings was largely an all-or-nothing affair managed from the Teams admin center. Admins could disable Copilot or Intelligent Recap across the entire tenant or for specific users, but they couldn’t let individual organizers decide in the moment without broad policy changes.

The Meeting AI toggle hands that power to the organizer, reducing the need for IT to field one-off requests or create complex, layered policies for different departments. That said, admins will still have overrides: they can likely prevent certain users from ever seeing the toggle if corporate governance demands it.

For compliance teams, the toggle provides a paper trail. Meeting participants and auditing tools will be able to see when AI was toggled, creating a clear record of which segments were or weren’t processed by Copilot. This could simplify e-discovery and data subject requests under regulations like GDPR.

Microsoft has not yet published detailed admin documentation, but expect a new meeting policy parameter—something like -AllowMeetingAIToggle—to appear in PowerShell and the Teams admin center by the time the feature ships.

How we got here: the push for user choice

The road to a live AI toggle has been paved by a rapid—and sometimes rocky—AI rollout in Teams. Here’s a brief timeline:

  • 2023–2024: Microsoft introduced AI-powered meeting recaps and the early Copilot in Teams, but controls were strictly admin-level. Users either had AI or they didn’t.
  • Early 2025: Copilot gained the ability to be invited to meetings rather than joining automatically, giving organizers a pre-meeting choice—but once the meeting started, that choice was locked.
  • Mid-2025: Customer feedback intensified, with large enterprises voicing concerns at Microsoft conferences. Many requested granular, in-meeting controls, especially for calls involving external parties or legal discussions.
  • Late 2025: Microsoft announced on the Microsoft 365 Roadmap that an in-meeting AI toggle was in development, though details were scarce.
  • January 2026: The company confirmed the toggle would include Copilot, Facilitator, and Intelligent Recap, and set a target of mid-2026.
  • May 2026: Roadmap updates pinned the rollout to July 2026, with the user interface described as a single toggle in the meeting toolbar.

This evolution mirrors a broader industry trend: as generative AI proliferates, users and regulators alike are demanding opt-out mechanisms that work in real time, not just at setup.

What to do between now and July 2026

If you’re an IT administrator or a business decision-maker, there’s groundwork to lay before the toggle arrives.

  1. Audit your licenses. Identify which users have Teams Premium or Copilot licenses today, and determine who should have organizer/presenter privileges. The toggle will only benefit those with the right license, so now is the time to adjust licensing to match business needs.
  2. Review your meeting policies. Take stock of how you currently manage AI features via the Teams admin center. Decide whether you’ll keep those tenant-level settings as a baseline and let the toggle act as a per-meeting override, or whether you’ll loosen existing restrictions once organizers have the toggle.
  3. Draft communication and training. Create simple guides for meeting organizers: when to use the toggle, what it does (and doesn’t) do, and how to avoid accidental AI exposure. Include screenshots once the feature reaches public preview or early rings.
  4. Consider security and compliance implications. Work with your legal and compliance teams to update internal policies. Document that meeting organizers now have the ability to prevent AI data capture, and clarify whether they are expected to use it for certain meeting types.
  5. Watch for preview builds. Microsoft typically rolls features to Targeted Release or public preview rings several weeks before broad availability. Enroll a test group to try the toggle and provide feedback.

For everyday users, the message is simple: come July 2026, if you host meetings that involve sensitive or off-the-record conversations, you’ll finally have a switch—right inside the meeting—to turn off the AI assistants that you’d rather not listen in.

The bigger picture: AI governance is maturing

The Meeting AI toggle is a small feature with large implications. It signals that even as Microsoft pushes AI deeper into the workplace, it’s willing to give power back to users—at least those with premium licenses. Expect more such controls to appear across Microsoft 365. Word, Excel, and PowerPoint already have Copilot sidebars that can be dismissed, but true real-time toggling during collaborative work might follow a similar pattern.

Competitors are watching. Zoom has offered meeting host controls for AI Companion since 2024, but the ability to switch features on and off during a live session is not universal. Google Meet’s AI notes can be started and stopped, but the integration is less comprehensive. Microsoft’s move could pressure the whole market to offer fine-grained, in-session AI governance.

In the end, the toggle is not about turning off AI forever—it’s about making AI optional, situational, and respectful of boundaries. For millions of Teams users, that’s a welcome evolution.