AI Agents in Action, Second Edition

Get the eBook free when you register your print book at Manning.

Build AI agents that do more than respond. Give them tools, connect them to knowledge, coordinate their work, and evaluate how well they perform.

AI Agents in Action, Second Edition teaches you how to design, build, evaluate, and deploy AI agents through working Python examples. You’ll progress from a simple agent to tool-using assistants, coordinated multi-agent workflows, and deployable systems. Along the way, you’ll learn how the pieces fit together—and where to look when an agent needs improvement.

What’s new and what’s included in the second edition:

• 95% rewritten, with new chapters and expanded examples to help you build, evaluate, and deploy today’s AI agents
• Understand how agents fit together with five dedicated chapters covering prompting and context, tools, reasoning and planning, knowledge and memory, and evaluation and feedback
• Hands-on Model Context Protocol (MCP) coverage, including connecting agents to MCP servers and building your own servers to make tools available to agents
• New chapter on evaluation and feedback: develop rubrics for LLM-as-judge evaluations, apply test-driven agentic development, incorporate human and automated feedback, and observe agent behavior with Phoenix
• New chapter on the agentic loop: explore its three layers and learn how to build task-focused agents and longer-horizon, goal-driven systems such as deep research agents
• New chapter on metacognition harnesses, the structures around an agent’s thinking, planning, and learning that help it monitor and adapt its approach
• A new chapter of practical tips and techniques for applications including RAG help desks and deep research agents
• Expanded multi-agent workflows, with assembly-flow, orchestration, and collaboration patterns for dividing complex work among agents
• Deeper coverage of reasoning and planning, explaining the internal agentic loop and how agents work through multi-step tasks
• Expanded knowledge and memory coverage, including hybrid search for retrieval-augmented generation (RAG) and practical ways to use memory in agents
• A dedicated chapter on voice-driven AI agents, including a complete containerized web platform that demonstrates real-world deployment patterns for multi-agent systems

This is more than a refresh of the first edition. It takes you from an agent’s first response to systems that use tools, work together, draw on knowledge, and incorporate feedback with practical examples throughout.

What's inside

• Autonomous agent design and deployment
• MCP-based tools, resources, prompts, memory, and server integrations
• Reasoning and planning patterns including ReAct, Reflexion, Tree-of- Thought, and Sequential Thinking

About the reader

For intermediate Python programmers. No experience with AI agents and agentic systems required.

About the author

Micheal Lanham is a software and technology innovator with over 20 years of industry experience. He has authored books on deep learning, including Manning’s Evolutionary Deep Learning.