How does Sembly identify speakers?

For online meetings:

• Sembly automatically identifies speakers by name, even if they are not registered in Sembly. Names will be taken from the conference platform. This feature is available for Google Meet, Zoom, Microsoft Teams, and WebEx. 

For audio recordings and audio uploads: 

• Speakers are recognized by voice and assigned as Speaker 0, Speaker 1, etc. Later, names can be changed to a real name in meeting settings.  
• To assign speakers by name, each speaker has to be a registered user and record a Voice ID sample. The voice sample will be matched to the speaker on the call and automatically assigned to its owner. 

Please note that speakers may be misidentified at times for various reasons. Here are some of them:

• Sembly labels speakers based on the information gathered from the meeting platform. That's why you can notice that parts of the meeting are already assigned to a particular speaker in the transcript. Sometimes, Sembly labels the speaker as an Unknown Speaker due to the following reasons: multiple speakers were talking at the same time, or the meeting platform did not send the speaker information to Sembly.
• Audio quality and conversation flow significantly impact the outcome. Please read this article to find out more.
• A speaker may be labeled as the Unknown Speaker (or Speaker 0, 1, etc.) when there's minimal speech that wasn't sufficient for Sembly to identify and label the speaker's name accurately in the transcription.

If the issue persists, we need to investigate it. We will need your consent: fill in the consent form with the meeting session ID and update us by email. Additionally, for the investigation, we need the following information:

1. Detailed description of what's wrong (E.g., Speaker John identified as Tom)
2. Examples: list of timings from transcription (E.g., 15:16, 26:07, 29:10)
3. Additional details (In free form, not required)