Question

A retail company runs a RAG-based support agent in Azure AI Foundry that grounds answers on product manuals indexed in Azure AI Search. Over the past week, agents report that recently uploaded manuals are not appearing in answers, even though the source documents exist in the connected blob storage. You need to diagnose whether the problem lies in the data ingestion and indexing pipeline before adjusting the retrieval logic. Which action should you take FIRST?

Accepted Answer

Review the Azure AI Search indexer execution history for failed or skipped documents and inspect indexer error and warning details.

Answer

Increase the vector search top-k value in the agent's retrieval configuration to return more candidate documents.

Answer

Switch the index from keyword search to hybrid semantic ranking to improve relevance of the missing documents.

Answer

Regenerate the embedding model deployment key and reconfigure the skillset connection in Foundry.