Agentic GraphRAG

Chapter 5. Agentic Graph-Based Tool Orchestration

A note for Early Release readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 6th chapter of the final book. The GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at @oreilly.com.

Intelligence without the ability to act is merely contemplation. An AI agent that can understand, reason, and plan but cannot interact with the world remains trapped in a prison of its own thoughts. This constraint drives the entire evolution of agentic systems: agents need tools as essential extensions of their intelligence.

Large language models (LLMs), at their core, perform a single operation: transforming text input into text output. They cannot query databases, send emails, or interact with any external system on their own. They are minds without hands. Tools are the bridge between an agent’s intelligence and the world it seeks to influence; with tools, an agent can transition from describing what should be done to actually doing it.

The breakthrough comes from treating tools as interconnected data in a knowledge graph rather than as hardcoded logic. The Model Context Protocol (MCP) makes this transformation practical. It provides a foundation for building software ...