I run a service called Dictanote that lets users dictate notes (think of Evernote + Nuance Dragon).
A repeated request I get from mainly medical professionals is if I could transcribe audio files. It would be cool if say someone can just drop audio files into their Dropbox, right click and select transcribe. Then Dropbox would use Amazon Medical Transcribe or some other API to get the transcript and add it back into user's Dropbox.