How to know who said what in a meeting

Knowing who said what in a meeting comes down to a transcript that matches the calendar invite to the voices. Say names out loud on the call, and correct whatever is left afterward. Wispr Notetaker labels speakers in your final transcript using calendar, dictionary, and conversation context.

Why speaker attribution is the part that breaks

A transcript can have your words right and still put tag them to the wrong person. If the quote lands under the wrong name, the follow-up goes to the wrong person. Someone gets credit for a commitment they never made, and the customer quote you wanted is unusable.

Three things make attribution hard:

• Several people are in the same room, so there are no separate windows or invite lines to tell them apart
• Names are never spoken, so there is nothing in the conversation to tie a voice to a person
• The meeting was never on a calendar, so there is no attendee list to match against

Most tools handle the easy case: a scheduled remote call where everyone is in their own window with their own audio. The gap shows up in the messy ones. Knowing which of those three problems you have tells you which fix below is worth your time.

Four ways to get who-said-what right

These run from free and manual to automatic. Most people end up combining the first one with the last one.

1. Say names out loud during the call. "What do you think, Dan?" gives any tool a clue it can use, and it gives you one too when you reread the transcript later. This is the single cheapest habit for a room with shared audio.
2. Use your platform's own transcript. Zoom, Google Meet, and Teams each keep their own transcript. Coverage depends on your plan and your admin's settings, and anything that happens outside that platform will not be in it.
3. Fix labels by hand after the meeting. Open the transcript, listen back at the points where speakers change, and relabel. It works on any recording and it is honest about who said what. It also costs you the length of the meeting again, which is why nobody does it twice.
4. Use a notetaker that labels speakers for you. A tool like wispr notetaker with access to the calendar invite and your own vocabulary can name people without you doing anything, then let you correct what it could not identify.

How Wispr Notetaker labels speakers

Wispr Notetaker records on your device instead of joining the call as a participant, so nobody sees a bot in the meeting. (Always tell everyone you're recording)

Here is the order it happens in:

1. Add the names and terms you use often to your personal dictionary
2. Start Wispr Notetaker on the call, in one click from the notification
3. Read the live transcript while you talk, with speakers shown as You and Them
4. Open the final transcript afterward and name anyone who was not identified

Wispr Notetaker reads your calendar, Slack, Gmail, and your personal dictionary before the meeting starts. That is where the names and the unusual terms come from: a colleague's name spelled right because they are on the invite, an acronym spelled right because it is in your dictionary. After the meeting it makes a second pass over the audio, and the final transcript comes back with the speakers labeled. If it cannot place a voice, name that speaker once and the label fills in everywhere it appears.

It works wherever your meetings happen, including Zoom, Google Meet, Teams, and a Slack huddle, and it can start on an ad-hoc call that was never on your calendar.

How to check attribution before you trust it

Do not take any tool's word for this, including ours. Run one real meeting and check the parts that are hard.

1. Put the meeting on the calendar with every attendee on the invite, and add the names and terms you care about to your dictionary first
2. Include one stretch where two people are in the same room on one microphone, and have them address each other by name at least once
3. Ask one question that names a person out loud, and one that does not
4. Open the final transcript and jump to the three moments you remember most clearly, then check the name on each
5. Pull one quote you would actually paste into an email and see whether it is usable as written

The shared-mic stretch and the quote you would paste are where attribution usually shows its limits. If those do not hold up, you now know exactly which habit to add on the next call.

Who should rely on automatic speaker labels

Automatic speaker labels are worth relying on if your after-meeting work depends on precise words with the right name against them. That includes anyone piping meetings into Claude or ChatGPT, where one wrong label spreads downstream, and anyone pulling customer quotes, like journalists and researchers. It also helps if you run back-to-back calls, since Wispr Notetaker switches over with you instead of capturing the wrong one.

Not for you yet if you are on Windows, which is coming soon, or on iPhone or Android, which are on the roadmap. Wispr Notetaker is on Mac today. It supports English today, with more languages coming soon. You can use it in other languages, but expect mixed results until dedicated support arrives. And Wispr Notetaker will not label a person in a shared room who is never named out loud, so that case still needs a habit or a quick correction from you.

Start with the meeting you have tomorrow

The habit worth keeping either way is saying people's names in the conversation. It costs nothing, it helps every tool, and it helps you when you reread the transcript in a month.

Record your next meeting and read the final transcript before you decide.

Frequently asked questions

Why does my transcript say "Speaker 1" instead of a name?

Because nothing tied that voice to a person. There was no calendar invite to match against, or the person was never addressed by name in the conversation, or several people shared one microphone. In Wispr Notetaker you name an unidentified speaker once after the meeting and that name applies across the whole transcript.

Can a notetaker tell people apart in the same conference room?

Partly. When people in a room share one microphone, their audio arrives on the same stream, so labels depend on conversation clues. Have people address each other by name during the meeting. Anyone in the room who is never referenced by name will need you to label them once afterward.

How do I get names spelled correctly in a transcript?

Two sources do most of the work: the calendar invite, which carries the attendees' names, and your personal dictionary, where you add the terms and names you use often. Wispr Notetaker reads both, plus Slack and Gmail, so unusual names and acronyms come through spelled the way you write them.

Can I see who is speaking during the meeting, not just after?

Wispr Notetaker gives you a live transcript while the meeting runs, with speakers shown as You and Them. Named speaker labels arrive in the final transcript after the meeting, once it has made a second pass over the audio.

I already use another notetaker. Do I lose my old transcripts?

No. Your existing tool keeps its history, and you can run Wispr Notetaker alongside it on the same meeting to compare the two final transcripts before you commit. If you are coming from Granola, one-click import brings your notes across.

Does Wispr Notetaker keep recordings of my meetings?

Your meetings are yours. Meeting audio is encrypted and kept only temporarily to create your transcript, let you resume a meeting, verify quality, and troubleshoot. After that limited period, it's automatically removed. Wispr Notetaker doesn't create voiceprints or biometric profiles of you or anyone on your call.