AI Note Taking for Teams Meetings: Setup and Review
Microsoft Teams has two built-in AI note-taking paths and accepts third-party apps. Here is how to choose, enable, and check your output before acting on it.
Updated 5 min read
AI note taking in Teams meetings works through two built-in options tied to your Microsoft 365 plan: plain transcription, which records speaker turns with timestamps, and Intelligent Recap, which adds topic chapters and suggested action items. Both notify all participants automatically. Tools like fenbs let you keep the verified decisions and tasks from the meeting in a shared record any connected AI can read in the next session.
The practical challenge is not enabling the feature — it is knowing which output is accurate enough to act on. A recap of a long meeting can turn "we should look at this" into an assigned action, or drop a qualifying condition that changes the meaning of a decision entirely. Review before you share.
The two built-in options in Teams
Plain transcription is available on more Microsoft 365 plans and produces a timestamped text record with speaker labels. Intelligent Recap — automated chapters, a full meeting summary, suggested action items, and a log of when your name appeared — requires Teams Premium or a Microsoft 365 Copilot license. Check Microsoft's support pages (opens in a new tab) for current plan eligibility, as requirements change with licensing updates.
- Transcription: captures every spoken word with speaker attribution, stored in the meeting chat and the organizer's OneDrive after the meeting ends.
- Intelligent Recap: generates topic chapters, a summary, suggested action items, and marks when your name appeared. Requires Teams Premium or Copilot.
- Meeting notes (manual): a shared document editable during the call, available on more plan types than Intelligent Recap.
If you do not see the recap option after a call, the meeting organizer's license typically controls which features are available for everyone in the meeting.
How to enable transcription and recap
- Ask your IT admin to allow transcription in the Teams admin center under Meetings > Meeting policies. Without this policy, the Start transcription option will not appear in meetings.
- During a meeting, click the three-dot menu (More actions) and select Start transcription. A notification appears for all attendees.
- If your license includes Intelligent Recap, starting a recording also triggers the recap. It generates automatically after the meeting ends.
- After the meeting, open the event in your Teams calendar or meeting chat and click the Recap tab. Transcripts and summaries typically appear within a few minutes to one hour.
- Review the suggested action items before distributing them. Remove any that do not reflect a real commitment made in the call.
Consent happens automatically: Teams shows a visible notification to all participants when transcription starts, so you cannot transcribe silently. For meetings with external guests, check your organization's recording policy and inform guests in the invite if required.
Third-party note-taking apps in Teams
If your plan does not include Intelligent Recap, or you want consistent notes across Teams, Zoom, and Google Meet, a third-party app may be worth adding. These tools either integrate through the Teams app store or join as a meeting participant. Before choosing one, check:
- Where transcript data is stored and who can access it.
- How participants are notified — most apps display a banner or play an announcement on entry.
- Browser versus desktop client compatibility, since some apps only work in the Teams desktop app.
- Accuracy on a short, low-stakes call before using the tool in a high-stakes meeting.
The AI meeting note taker guide compares options across platforms. fenbs does not connect to Microsoft Teams or import meeting transcripts directly.
What to check in the recap before you act
A solid review of a one-hour recap takes five to ten minutes. Look for these common errors before sharing or acting on the output:
- Speaker misattribution: a statement given to the wrong person, which can flip responsibility for an action.
- Proposals logged as decisions: "we should consider" recorded as a confirmed next step.
- Missing conditions: a decision noted without the qualifier that makes it applicable.
- Dropped context: an action item with no record of why it was assigned, making it easy to misread or skip.
Correct these before sharing. One wrong action item sent to a colleague creates more work than the recap saved. The AI output is a draft; it becomes a useful record only after a person confirms the key facts.
What the recap does not carry forward
Even an accurate recap stays inside Teams, OneDrive, or the third-party tool that made it. A developer asking an AI assistant about a project two weeks after the meeting will not see what was decided. The decisions and agreed tasks need to reach the place where the work is tracked.
Most teams copy action items into a task list and archive the rest. That works for simple tasks, but misses the decisions — the rules and reasoning that explain why a task is set up the way it is. Those are exactly what an AI assistant needs to avoid re-asking the same questions.
Keep what the meeting actually decided
After reviewing the recap, add confirmed decisions to fenbs as Decisions and rules, and verified action items as tasks on the relevant project. Every connected AI reads those before it works. fenbs does not import transcripts — you add checked conclusions directly, which means only verified information goes into the shared record.
Create a free accountFree for one person · No credit card · Works with Claude, ChatGPT, Cursor and more
How to do this with fenbs
- Create a free account and open or create the project the meeting covered.
- Run the Teams AI recap or transcription. Review the output for speaker errors, mislogged decisions, and missing conditions.
- Add confirmed decisions to fenbs as Decisions and rules, with the date and who decided. Add verified action items as tasks in the right lane.
- If the meeting changed an earlier rule or decision, update that record in fenbs so the next connected AI reads the current version.
- Connect your AI apps at Integrations. Each one reads the updated project context — including the decisions from the call — before starting work.
Teams that do this after every meeting find that the next AI session starts with a useful record instead of a blank page. The AI memory for teams guide covers the broader approach to shared records across a whole team.