Meeting detail page

Work with a finished meeting: summary and transcript, naming speakers, chatting with the meeting, and using follow-up assistants

Click a meeting in the overview to open its detail view. This also works for failed or cancelled bot meetings. This page is the control panel for everything you do with the meeting afterwards.

Header and actions

At the top you see the meeting name and status. Which other details and actions are available depends on the existing result and your permissions. Available meeting recordings offer these main actions:

  • Chat with meeting: opens a new chat with the transcript attached
  • Follow-up: quick access to pre-configured assistants that process the meeting in one click (see Follow-up assistants)
  • Create Word protocol: generates a formatted Word document to download
  • Share (icon): share the meeting with teammates
  • Delete (icon): remove the meeting along with the transcript and summary

Tabs: Summary and Transcript

Summary and transcript have their own tabs. For bot meetings without a saved transcript, the detail view initially shows the failure reason and possible next steps.

  • Summary: the generated summary with the components you selected when you created the meeting (Executive Summary, Participants, Agenda, Action Items).
  • Transcript: the full transcription text with speaker labels.

Use Copy to copy available text. Further actions such as Regenerate and Name speakers depend on the processing state and your permissions.

While a summary is being regenerated, the previous text stays visible and can be copied. A notice indicates that processing is in progress. You cannot start another regeneration while it is running. Placeholders appear only when no summary is available yet.

Recover failed recordings

Open a failed or cancelled bot meeting from the overview. The detail view explains the reason. If you own the meeting and automatic recovery is possible, choose Start recovery. Processing runs in the background; you can leave the page and reopen the same entry later.

Available content remains accessible. You can read and copy a displayed partial transcript. Editing and sharing are not available in this partial transcript view.

If the check finds that the recording is missing or a transcription service limit has been exceeded, automatic recovery is not possible. Follow the displayed guidance.

If the recording contains no recognisable speech, the detail view says so. This happens, for example, when the bot joined a meeting in which nobody spoke. Recovery would produce the same result and is therefore not offered.

If the transcription service response is lost, MeinGPT looks for the job already submitted. While acceptance remains uncertain, the recording is not submitted again. After automatic attempts are exhausted, recovery checks for the existing job again.

For older recordings, it may be unclear whether a transcription was already created. Automatic recovery is then unavailable. Support must first verify that the retained recording and an existing transcript belong together, so no new transcription is submitted.

Only the person who owns the meeting can start recovery. Workspace admins cannot trigger it for other people's meetings; ask the owner to open the meeting themselves. If no button appears, or the notice refers you to support (for example for a transcription service limit, rejection, failure or timeout, or a job that support has to check), contact support. Include the meeting name, date and time, your workspace, and the email address of the person who owns the meeting.

Naming speakers

MeinGPT detects when a different person speaks in every recording and separates the speaking shares (Speaker 1, Speaker 2, …). Whether a name appears automatically depends on how the meeting was created:

  • Bot in an online meeting: The names come from the meeting itself (participant display and the indicator of who is currently speaking), not from the voice. You don't need to do anything. If a voice can't be matched unambiguously, it stays Speaker N. MeinGPT doesn't guess.
  • Audio/video upload and live recording: There is no participant list here. Name each speaker once by hand (per speaker, not per utterance), then regenerate the summary.

Attention

A name exists only where the meeting shows a participant. A meeting room or a dial-in counts as one participant with one name, no matter how many people sit in front of it. If Teams runs during an in-person meeting, remote attendees are named correctly, but the people in the room stay unnamed or carry the name of the room device. Review the transcript before you share it.

There are two ways to rename them, both in the Transcript tab:

  • Name speakers in the action bar opens the list of all speakers. For each one you see Speaking share (as a percentage) and the number of utterances.
  • Clicking a speaker name above an utterance opens the same list, with that speaker's field focused.

In the name field you can type freely or pick a member of your workspace. For a member, their profile picture then appears next to their utterances.

Speaker names are also used in the summary and in any later chat attachment of the transcript. If you change them after the summary was generated, a hint says so and offers Regenerate right there.

Chat with the meeting

Once a meeting reaches Done (or Completed in the overview), you can talk to its content directly.

  • On the meeting detail page, click Chat with meeting. A new chat opens with the transcript attached. Ask any follow-up questions, request a summary, or extract action items.
  • In a new or existing chat, open + → Meetings, or type @ in the input area and choose Meetings. Search for the meeting and check its date and time before selecting it. The selected meeting appears as an attachment in the input area and is sent with your next message. Use × to remove it before sending.

Follow-up assistants

Next to Chat with meeting sits the Follow-up button: pre-configured assistants you apply to the meeting in one click – for example to extract action items or draft a follow-up email. One click opens a new chat with the chosen assistant and the transcript attached.

They are ordinary assistants – "Follow-up" is just the quick access. Workspace admins choose which ones (up to eight) appear here under Settings → AI & features → Meetings.

Automatic follow-up emails

After a meeting recorded by the notetaker, every invitee can automatically receive a summary by email – without a confirmation step.

Admins choose who receives it under Settings → Meetings → Follow-up emails after meetings: Off, Meeting owner only, Internal participants only or All participants. The default is off. Internal means an address on a verified domain of your organization. Enabling it never sends emails for past meetings.

Everyone can narrow this for their own meetings under Meeting settings → Follow-up emails – for example to Internal only, Only me or Off – but not widen it.

The person who records the meeting always receives the summary. Further recipients are the invitees of the calendar event, including those who did not join; meeting rooms receive nothing. Each person receives an individual email with the same summary: the topics discussed, decisions, next steps and open questions. The email says who shared the meeting, and replies go straight to that person. It contains no link to the recording and grants no access to it.

As the meeting owner, open the Follow-up email tab to see the sent content, the recipients and their delivery status. If the transcript is not enough for a reliable summary, nothing is sent and the tab shows why.

Recipients can use Email preferences for this organization in the email to stop further summaries from that organization. Opening the link alone does not unsubscribe anyone.

Related pages

Was this page helpful?