There are three scenarios which our Speaker Auto-Fill feature can help. Please note this feature works best with single speaker files. We continue to improve our speaker recognition technology for multiple speaker files, but it is still not perfect.

1. Single speaker files

If your file only has one speaker then Speaker Auto-Fill works great. 

  1. In the transcript, add the name of the speaker
  2. Select the "Speakers" tab in the top nav
  3. Click "Label Remaining"

2. Files with two speakers

We are still improving our speaker identification technology: It still not perfect. For that reason the Speaker Auto-Fill works a little differently for two-speaker files. It can help save some time, but you you'll still have to label every exchange for one speaker manually.

  1. In the transcript, manually add the name of the first speaker to all exchanges associated with that speaker
  2. In the transcript, add the name of the second speaker to one of the exchanges associated with that speaker
  3. Select the "Speakers" tab in the top nav
  4. Click "Label Remaining" for the second speaker

3. Files with more than two speakers

We are still improving our speaker identification technology: It is still not perfect. For that reason the Speaker Auto-Fill can help save some time, but you you'll still have to label many of the exchanges manually.

  1. In the transcript, manually add the names of all but one speaker to all exchanges associated with those speakers
  2. In the transcript, add the name of the last speaker to one of their exchanges 
  3. Select the "Speakers" tab in the top nav
  4. Click "Label Remaining" for the last speaker

For files with more multiple speakers, it is best to manually fill the speakers with the least number of exchanges. That way you can "Label Remaining" for the remaining speaker with the most exchanges and save more time.

Lastly, if you do something incorrectly and want to revert back, you can always click the UNDO button in the top nav.

Did this answer your question?