A call-center operator wants to build a workflow in Azure AI Foundry that takes recorded customer calls, converts the audio into text, and then has a language model summarize each call for the account notes. Which combination should the workflow use?
Select an answer to reveal the explanation.
Short Explanation
Audio and text are just different kinds of data, and most chat models are built to read and write text, not to listen to a recorded call directly. So the first job here belongs to a dedicated speech service, the kind built specifically to turn spoken audio into a written transcript. Once you've got that transcript as plain text, that's when the language model can step in and do what it's good at, condensing it down into a tidy summary for the account notes. A tool built for reading text out of images doesn't have anything to do with sound, and a safety filter is there to screen content, not transcribe it. Two different jobs, two different tools, chained one into the other, is what actually gets you from a recorded call to a written summary.
Full Explanation
The correct answer is D. Azure AI Speech converts the recorded audio into a text transcript, which can then be handed to the deployed language model to produce a concise summary, matching the two distinct capabilities this workflow actually needs. Option A is incorrect because the chat models typically deployed for this kind of summarization task work on text input, not raw audio waveforms, so passing audio directly would not produce a usable transcript or summary. Option B is incorrect because Azure AI Vision's OCR capability extracts text from images of printed or handwritten content, such as scanned documents or photos; it has no mechanism for processing audio signals. Option C is incorrect because the content safety filter's job is to screen text for policy violations, not to transcribe audio into text, so it cannot produce the transcript this workflow depends on. Chaining a dedicated speech-to-text service into the language model is the combination that actually matches what audio-to-summary requires.