Question 17
Domain 4: Implement Natural Language Processing SolutionsA user says to your voice assistant: "Set an alarm for tomorrow morning." The assistant correctly transcribes the speech but doesn't understand the intent. The assistant uses Azure Speech for STT and needs intent extraction. What should you integrate?
Correct answer: B
Explanation
Azure Speech handles speech-to-text, but intent extraction requires a language understanding service. CLU (Conversational Language Understanding) is used for intent recognition, so it can interpret "Set an alarm for tomorrow morning" as an alarm-setting request rather than just transcribing the words.
Why each option is right or wrong
A. Azure AI Language — Key Phrase Extraction
B. CLU (Conversational Language Understanding) — connected to Azure Speech Intent Recognition
Azure Speech only performs speech-to-text; it does not infer user intent from the transcript. To extract the meaning of an utterance like a request to create an alarm, you must connect Speech to a language-understanding model via Azure Speech Intent Recognition, which uses CLU (Conversational Language Understanding) for intent classification and entity extraction.
C. Custom Question Answering
D. Azure Translator with language detection