Home · Business · Admin & Office · Minutes & transcription
As of 13 August 2026, AI can transcribe your team meeting.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ityou
What the alternative costsA purpose-built alternative is Fathom, described as a free-first AI notetaker for Zoom, Meet and Teams.
If this goes wrong, names, words or decisions can be misheard and entered into the team's record until someone checks them against the audio.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Check your organisation's recording policy and obtain the required agreement from everyone attending before recording or uploading the meeting.
- Open the original recording and make sure it has clear audio, then note the meeting title, date, participant names and any specialist terms that may be misheard.
- Upload the recording to a chatbot or a meeting transcription tool and paste the prompt, including the participant list and any relevant terminology.
- Ask the tool to produce the full transcript rather than only a summary, and keep the original recording available for comparison.
- Search the transcript for every [inaudible], [uncertain] and speaker label, then replay each matching section and correct the text or attribution.
- Compare names, figures, dates, decisions and action points against the audio and your meeting notes before saving the checked transcript.
- Send the corrected transcript only to the people permitted to receive it, and store or delete the recording and transcript according to your organisation's policy.
Prompt
Transcribe the attached team meeting recording in full. Use speaker labels where you can identify speakers, and add timestamps at natural intervals and whenever the speaker changes. Preserve the speakers' wording rather than summarising or correcting it. Mark unclear audio as [inaudible] and uncertain words as [uncertain: ...]; do not guess. Keep names, figures, dates, technical terms and action points exactly as heard, and flag any passage that needs checking against the recording. Use these participant names only to help identify speakers: [PASTE NAMES AND ROLES]. Return the transcript in a clean format suitable for checking and sharing internally. Do not create minutes, infer decisions or attribute a statement to a person unless the recording supports it.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- Cannot reliably identify speakers when people talk over one another, use poor microphones or leave the camera layout unclear.
- Cannot recover words that are missing from the recording, so it marks gaps or guesses unless you instruct it not to.
- Cannot know whether a phrase is an internal name, product term or technical expression without context from you.
- Cannot obtain consent, decide who may access the transcript or take responsibility for retaining sensitive meeting content.
- Cannot replace checking important names, figures, commitments and decisions against the original audio.
Even on a YES, the friction has a name: consent and privacy and verification cost.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 2 |
| Total | 8 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI transcribe a meeting recording?
- Yes. Current meeting tools can produce a full transcript with speaker labels and timestamps from a clear recording, but you still need to check unclear passages, names and decisions against the audio.
- Is it legal to record a work meeting in the UK?
- That depends on consent, your organisation's policy, the purpose of the recording and the personal data involved. Check your employer's rules and tell participants clearly before recording or uploading the meeting.
- Can AI tell who said what in a meeting?
- Often, but not reliably when speakers overlap, microphones are poor or several people sound similar. Give it the participant list and verify every important attribution against the recording.
- What is the best AI tool for transcribing meetings?
- Fathom, Fireflies.ai, Notta, Otter.ai and tl;dv are purpose-built meeting transcription tools in this category. Choose one that works with your meeting platform and meets your organisation's recording, storage and access requirements.
Nearby answers
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.