How to Transcribe a Conference Call (Multi-Party Calls)
Learn how to transcribe a conference call with multiple speakers: record it, label participants, fix crosstalk, and turn the transcript into clean action items.
A conference call is one of the harder things to transcribe accurately, because several people speak, sometimes at once. The reliable approach: record the call, upload the audio to a transcription tool, turn on speaker labels to separate each participant, then proofread the spots where voices overlap. From there you can pull out clean notes and action items.
This guide walks through the full workflow, including the multi-speaker accuracy realities most tools gloss over.
1. Record the conference call
You can't transcribe what you didn't capture, so start with a clean recording.
- Zoom / Microsoft Teams: Use the built-in record button. After the call, download the audio (or video) file from your recordings. For Teams specifically, see our walkthrough on how to transcribe a Teams meeting.
- Softphone / dialer: Many business dialers (the apps sales and support teams use) have a record toggle. Enable it before dialing.
- Phone bridge / conference line: Dial-in bridges from providers like RingCentral or a PBX usually offer a recording option in the host controls.
- Fallback: If your platform can't record, a separate recorder app on a phone placed near a speaker can work, though room audio tends to be noisier and lowers accuracy.
Whatever the source, aim for the highest quality export available. Each participant on a separate, clear channel is ideal; a single muffled room mic is the worst case for multi-speaker accuracy.
2. Upload, label speakers, and export
Once you have an audio file, the transcription itself is straightforward.
- Upload the file. TranscribTxt runs on ElevenLabs Scribe and supports 99 languages, so most common conference-call formats and languages are covered.
- Enable speaker labels. This is the step that separates participants. Speaker labels (speaker diarization) are available on the Pro and Business plans. The tool detects when the voice changes and groups the text by speaker, so you get distinct blocks instead of one undifferentiated wall of text.
- Review and rename. The tool outputs generic labels like Speaker 1 and Speaker 2. Replace them with real names once you confirm who is who, usually obvious from the first few lines.
- Export. Download the transcript as TXT for notes, SRT for captions, or JSON if you're feeding it into another system.
Audio is deleted after transcription, which matters for sensitive business calls.
3. Multi-speaker accuracy realities
This is where conference calls differ from a single-person voice memo, so it's worth being honest about what to expect.
Diarization is pattern detection, not magic. Speaker labeling works by recognizing changes in voice characteristics. It tends to do well when people take turns and poorly when they don't. To understand how the underlying technology decides who said what, read speaker diarization explained.
Crosstalk is the main enemy. When two people talk over each other, the audio physically blends, and no tool can cleanly split overlapping words. Expect the roughest patches of any transcript to be the moments where participants interrupted one another.
Tips to improve multi-speaker results:
- Ask participants to use headsets rather than laptop speakers; this reduces echo and bleed.
- Encourage one-speaker-at-a-time discipline, or at least clear handoffs.
- Mute participants who aren't speaking when the platform allows it.
- Prefer separate channels per participant if your platform offers them.
- Record at the source rather than re-recording room audio.
Even with good audio, treat the first draft as a draft. Accuracy depends on accents, microphone quality, background noise, and how disciplined the conversation was. Always proofread anything important, especially the speaker attributions, before you act on the transcript.
4. Turn the transcript into notes and action items
A raw transcript is rarely the deliverable; the notes are. With a clean, speaker-labeled transcript, this gets much easier.
- Scan by speaker. Because the text is grouped by participant, you can quickly see who committed to what.
- Pull decisions and owners. For every action item, capture the task, the owner, and any due date mentioned. Speaker labels make ownership unambiguous.
- Summarize at the top. Lead your notes with a short summary, then list decisions, then action items, then open questions.
If you want a repeatable structure for this, follow our guide on how to write meeting minutes from a recording. The speaker-labeled export from step 2 plugs directly into that process.
5. Get consent before you record
Recording a multi-party call carries legal obligations that vary by location.
Some jurisdictions follow one-party consent, meaning a single participant (you) can record. Others require all-party consent, where every person on the call must agree. Rules differ by state and by country, and a call can span multiple jurisdictions when participants are in different places.
Practical guidance:
- Verify the law that applies to every participant's location before recording.
- When uncertain, default to announcing the recording at the start of the call and giving people a chance to object.
- Keep recordings secure and delete them when no longer needed.
For a deeper look at the considerations, see do you need consent to record and transcribe a meeting. Note that this article is general information and not legal advice; consult a qualified professional for your specific situation.
Plans and pricing
TranscribTxt offers a Free plan with 5 files per month and no card required, so you can test multi-speaker transcription on a real call first. Speaker labels unlock on the paid tiers: Pro is $12/mo for 1,200 minutes, and Business is $29/mo for 6,000 minutes. Both Pro and Business include speaker labels and export to TXT, SRT, and JSON.
Wrapping up
To transcribe a conference call well: record cleanly at the source, upload and turn on speaker labels to separate participants, accept that crosstalk will need manual cleanup, then convert the labeled transcript into structured notes. Handle consent up front, and you'll have an accurate, usable record of even your busiest multi-party calls.
Frequently Asked Questions
How do I transcribe a conference call?
Record the call through your dialer, Zoom, Teams, or a phone bridge, then upload the audio file to a transcription tool. Enable speaker labels to separate participants, review the draft for crosstalk errors, and export the finished transcript as TXT, SRT, or JSON for your notes.
How do you transcribe a call with multiple people?
Use a transcription service with speaker diarization, which detects voice changes and groups text by speaker. Clear audio and minimal crosstalk improve results. After transcription, label each speaker by name, then proofread sections where people talked over each other, since overlapping speech is the hardest part for any tool.
Can I transcribe a Zoom or Teams conference call?
Yes. Record the meeting in Zoom or Teams, download the audio or video file, and upload it to a transcription tool. Both platforms can also generate native captions, but a dedicated transcriber typically gives cleaner exports and more flexible speaker labeling for multi-party business calls.
Do I need consent to record a conference call?
It depends on your jurisdiction. Some regions use one-party consent, others require all parties to agree. Laws vary by state and country, so verify the rules that apply to every participant's location before recording. When in doubt, announce the recording at the start. This is general information, not legal advice.
How accurate is conference call transcription?
Accuracy depends heavily on audio quality, accents, and how often people talk over each other. Clean, single-speaker-at-a-time audio can transcribe well, while heavy crosstalk lowers results. Speaker diarization helps organize multi-party calls, but you should always proofread important transcripts before relying on them.