Feedback
Edited
Is feedback used to train Nabla or for troubleshooting purposes, or both?
Both.
Feedback audio is completely de-identified using a "beep" on the audio in place of PHI.
Why is just the last 2 minutes of audio captured? Can audio from earlier in a string also be included in feedback?
Only the last 2 minutes of audio is captured to adhere to privacy and only a short snippet is attached when audio is shared.