AI-900 Practice Question: Describe Artificial Intelligence workloads and considerations

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure AI Speech text-to-speech

  Why it's wrong here

  Text-to-speech synthesizes spoken audio from text, which could voice an answer but cannot determine what the answer should be. On its own it has no knowledge retrieval capability, so it cannot find warranty or return-policy content in the company's documents and would require another service to supply the response text first.

• ✗

  Azure AI Custom Vision image classification

  Why it's wrong here

  Custom Vision trains image classification or object detection models on labeled pictures, for example identifying defective products in photos. It works with visual data rather than text documents, so it cannot extract return-policy answers or power a conversational agent that responds to written customer questions about warranties.

• ✗

  Azure AI Vision spatial analysis

  Why it's wrong here

  Spatial analysis in Azure AI Vision interprets video streams to detect people's presence and movement in physical spaces, such as counting occupancy in a store. It produces no textual answers and cannot read policy documents or respond to warranty questions, so it is unrelated to building a document-grounded question-answering agent.

• ✓

  Azure AI Language question answering, grounded on the company's policy documents

  Why this is correct

  Question answering in Azure AI Language builds a knowledge base from supplied documents such as policy PDFs and returns precise answers to natural-language questions. Grounding responses in the existing documents means staff do not hand-author Q&A pairs, and the agent answers only from approved content, which fits the warranty and returns scenario exactly.

Go deeper

Related to this question