The audio transcription process starts when a call is answered at the agent desktop, and the Voice AI desktop app is connected. Then, the transcript data is returned through the data set, processed, and the output is sent to the agent.

Begin your troubleshooting by verifying whether the Voice AI desktop app is connecting successfully when an interaction starts. If a connection is established, the browser console should start showing events on the agent desktop. If the console does not show any events, verify that the data flow that retrieves transcripts and the queue processors that process transcripts are running.