Question 9
Domain 2: Implement Knowledge Mining and Azure AI Search SolutionsA financial analyst needs to find documents that are conceptually similar to a given research report, not just keyword-matching. The solution must scale to 2 million documents. Which Azure AI Search feature should you use?
Correct answer: C
Explanation
Vector search with embedding fields finds documents by semantic similarity rather than exact keywords, using vector representations of the research report and documents. Azure AI Search supports this for large-scale retrieval, making it suitable when the system must "find documents that are conceptually similar" across 2 million documents.
Why each option is right or wrong
A. Full-text search with Lucene proximity queries
B. Semantic search with semantic configuration
C. Vector search with embedding fields
Azure AI Search vector search is the feature designed for nearest-neighbor retrieval over embeddings, which is what you need when the query is a research report and the target is conceptual similarity rather than term overlap. In Azure AI Search, vector fields are indexed and queried using similarity metrics such as cosine or dot product, and the service is built to handle large indexes at scale, including millions of documents, so 2 million documents is well within the intended use case.
D. OData filter with category fields