Part VI

Part VI: Agentic AI

Building autonomous AI agents that use tools, plan multi-step tasks, and collaborate in multi-agent systems.

Chapter opener illustration: Part VI: Agentic AI.

"We build our tools, and then our tools build us."

CompassCompass, Goal-Seeking AI Agent

Part Overview

Part VI covers the complete landscape of agentic AI: from single-agent foundations through tool use and protocols, multi-agent orchestration, and specialized agents for code, browsers, and research. You will build agents that perceive, reason, and act autonomously; connect them to tools through standardized protocols (MCP, A2A); compose them into multi-agent teams; and design specialized agents for code generation, web automation, and data analysis. The part closes with a Tools of the Trade chapter on the agent stack (LangGraph, CrewAI, AutoGen, smolagents, PydanticAI).

Chapters: 5 (Chapters 26 through 30). These chapters draw on every preceding part and prepare you for the multimodal, evaluation, and production topics that follow. Agent safety and security gets its own dedicated treatment in Chapter 49 within Part X.

Big Picture

Moving beyond single-turn question answering, Part VI covers the rapidly evolving field of AI agents: systems that plan, use tools, collaborate with other agents, and take actions in the real world. This part connects everything you have learned so far into autonomous, production-ready systems.

What's Next?

This part begins with Chapter 26: AI Agent Foundations. Each chapter builds on the previous one, so we recommend reading Part VI in order.