Question

You are building an Azure AI Foundry agent that must answer employee questions using your company's internal policy documents. The documents contain thousands of PDF pages that are updated monthly, and answers must cite the specific source passages to reduce hallucinations. Which approach should you implement to ground the agent's responses?

Accepted Answer

Configure a retrieval-augmented generation (RAG) pattern using Azure AI Search as a vector index knowledge source connected to the agent.

Answer

Fine-tune a base LLM on the policy documents each month and deploy the updated model to the agent.

Answer

Add the entire document set into the agent's system prompt so the model can reference it directly.

Answer

Increase the model's context window and rely on the model's pretrained knowledge to answer policy questions.