AI-103 Practice Question 19: Fast Transcription Choice
Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 2 · question 9 of 10
A newsroom uploads finished 40-minute interview recordings and an editor waits at the keyboard for the transcript before cutting subtitles. Nothing is live; each file is complete before processing begins. The team's current batch transcription jobs sit in a queue and the editor sometimes waits ten minutes for a job that has not started. They want the transcript returned in the response to the request that submits the file, produced faster than the audio's own duration.
Which Azure AI Speech capability fits?
Show answer and explanation
Answer:
Fast transcription transcribes an audio file and returns the results synchronously, faster than real time; the transcript comes back in the response to the call that posted the file, which is what the editor is waiting on. It supports diarization, multi-channel audio, language identification and word-level timestamps; the trade-off is that it returns the display form of the text only, and the file must sit inside the documented size and duration limits.
- A is wrong because real-time transcription is built for streaming audio as it is spoken. It emits partial results as the audio arrives, so a 40-minute file takes at least 40 minutes to push through, slower than the audio, not faster. - B is wrong because batch transcription is asynchronous by design: you submit a job, it is queued, and you poll for completion. Polling more often does not make the job start sooner, and batch is aimed at large volumes of stored audio, not at a person waiting. - D is wrong because Content Understanding analyzers process content asynchronously and are aimed at extracting structured fields from media. Pulling in a second service to obtain a transcript Speech already returns synchronously adds cost and latency.
Reference: Use the fast transcription API