AI-103 Practice Question 36: AI Search Tool Existing Index
Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 4 · question 6 of 10
A company already runs a production Azure AI Search index over its knowledge base. Two other applications query it, an indexer keeps it current on a schedule, and the relevance has been tuned with hybrid queries and semantic ranking. A new Foundry agent must answer from that same content, and the search team insists there be exactly one copy of the content and one place relevance is tuned.
How should the agent be grounded?
Show answer and explanation
Answer:
Microsoft's own guidance on the file search page lists the alternatives, and the first row is this case: to "search existing Azure AI Search indexes," use the Azure AI Search tool. The agent queries the index the search team already owns, so there is one copy of the content, one indexer schedule, and one place where hybrid and semantic relevance is tuned.
- A is wrong because file search runs its own ingestion: parsing, chunking, embedding into a vector store. That produces a second copy of the content with its own freshness lag and its own relevance behaviour, which is precisely what the search team ruled out. - C is wrong because web search grounds on the public web. It cannot see the internal knowledge base, and public documentation is not the same corpus. - D is wrong because it is A with an extra moving part: a nightly export job, a rebuild, and a window in which the agent's answers are a day behind the applications querying the live index.