You are integrating knowledge for multi-agent consumption: product manuals in Azure AI Search, live inventory via an MCP tool, and policy PDFs requiring semantic retrieval. Which knowledge integration design is most appropriate?
Select an answer to reveal the explanation.
Short Explanation
Different knowledge needs different channels. You route manuals and policies through semantic/hybrid RAG, hit live inventory with MCP tools, and keep source-specific access controls. Public web-only, hourly prompt stuffing, and filename-only matching miss the mark.
Full Explanation
Knowledge integration for multi-agent consumption includes search, RAG, MCP sources, and semantic search. Routing document Q&A through semantic/hybrid RAG, live inventory through MCP tools, and source-specific access controls assigns each knowledge type to the right access pattern with security preserved. Public web only loses private corpus fidelity and governance. Hourly inventory prompt stuffing confuses operational data with prompt memory and scales poorly. Filename-only matching underpowers policy retrieval. Exam tip: RAG for documents; tools/MCP for live systems of record.