Agentic Context Engineering: The Multi-Source Query Problem

~13 min read

The book's concrete motivating example — a single query spanning Linear, Calendar, Gmail, and Slack — shows exactly why naive RAG breaks down, and introduces the 3-layer agentic solution.

Agentic Context Engineering: The Multi-Source Query Problem is a Pro topic

Sign in, then upgrade to Pro or Power to unlock this topic and the full AI Engineering curriculum.

Key points

  • The book's motivating example: 'What's blocking the Chicago office project, and when's our next meeting?' requires Linear, Calendar, Gmail, AND Slack — no naive RAG setup can handle this
  • The hard part isn't that the answer spans multiple documents (ordinary RAG handles that) — it's that it spans multiple SYSTEMS with entirely different data shapes
  • The book's solution: build an Agentic context retrieval system with three critical layers, making dynamic per-query decisions rather than using one fixed pipeline
  • The three named layers are Ingestion (getting differently-shaped sources in and fresh), Retrieval (searching the right sources the right way), and Generation (a trustworthy, verifiable response)
  • The book is candid this is a substantial undertaking — 'months of engineering before your first query works' — not a small tweak on manual RAG