Rather than the recording be a single file, ideally the transcription would be per user and be multiple files.