Agreed!
One suggestion: would it be possible to give the transcriptionist context that would be useful for more accurate transcriptions? For example, it seems to me that the transcriptionist should see the (first) name of both the caller and the recipient, if available. That way, when the caller leaves a voicemail starting "Hi Peyton, this is Kritika," it can be transcribed accurately. (I usually get "Hi (Pain?) this is (critical?)" or something worse than that!)
But overall, again, awesome.
-Peyton
Author