When diio analyzes in-person meeting audio, it identifies the different voices in the recording and labels them as "Voice 1", "Voice 2", etc. For the analysis to be correct — commitments attributed to the right person, mood by participant, speech metrics — you need to assign each voice to their participant.
When to assign voices
You can do it at two moments:
During upload via WhatsApp. diio asks directly in the chat which voice corresponds to each participant. This is the most convenient moment.
Later, from the detail view. If you didn't do it during the upload or want to correct an assignment, you can do it from the Participants card in the meeting detail view.
How to assign voices
diio shows audio fragments or transcript for each detected voice.
For each voice, select the participant's name from the list.
Confirm. The analysis is automatically recalculated with the correct names.
What happens if you don't assign voices
The analysis is generated anyway, but commitments and metrics remain attributed to "Voice 1" or "Voice 2" instead of to the actual participants. This makes follow-up difficult and can generate errors in CRM sync.