No. 426 of 500 No. 52 in Generative AI & LLMs

AI Agents in Action

Build, orchestrate, and deploy autonomous multi-agent systems

By Micheal Lanham

Published
July 14, 2026
Print length
391 pages
Reading level
Intermediate
ISBN-13
978-1633434530
Publisher
Manning Publications
Written for
Intermediate Python developers who want to prototype single-agent and multi-agent LLM systems with working code.

Synopsis

Micheal Lanham's Manning guide teaches developers to build autonomous AI agents and connect them into multi-agent systems. It opens by defining agents and their natural-language interfaces, compares open-source and commercial models, and covers prompt engineering before moving to OpenAI's GPT Assistants with custom actions and file uploads. From there the book builds multi-agent systems with Microsoft's AutoGen and AutoGen Studio and with CrewAI, then introduces behavior trees, a control structure familiar from game AI, as a way to orchestrate agents predictably. Later chapters cover retrieval-augmented generation and memory, advanced prompting methods such as chain of thought and tree of thought, and planning and feedback loops, with Microsoft's Prompt Flow and the author's open-source Nexus platform as recurring tools.

Each chapter extends working code from companion GitHub repositories, and the examples run against both hosted APIs and local models. Readers gain a practical sense of the components every agent needs, including personas, actions and tools, memory and knowledge, reasoning and planning, and evaluation, and they see several orchestration styles side by side. A June 2026 second edition rewrote more than 95 percent of the text around the Model Context Protocol, the OpenAI Agents SDK and agent evaluation, so readers buying today should check which edition they receive.

Scorecard

Editorial score

How it scored

Editorial score: 59.5 / 1005 weighted criteria

  • Influence30%

    16.5/30
  • Authority25%

    15.0/25
  • Practical value20%

    14.0/20
  • Currency15%

    7.5/15
  • Readability10%

    6.5/10
  • Editorial score

    59.5/100

How the rank was computed

Rank No. 426 of 500 comes from a composite of 49.0: observable signals 22.2 of 55 points plus editorial review 26.8 of 45 points.

All 11 inputs
InputPoints
Observable signals22.2/55
Reach (Amazon ratings volume)6.4/16.5
Consensus across best-of lists5.5/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted5.9/8.3
Amazon sales rank1.6/5.5
Recency2.8/2.8
Editorial review26.8/45
Author authority6.8/11.3
Influence7.4/13.5
Practical value6.3/9.0
Readability2.9/4.5
Currency3.4/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A hands-on tour of 2024-era agent frameworks, strong on progressive code projects but tied to tools that are already changing.

Strengths

  • Progressive projects that build on each other
  • Covers several frameworks side by side
  • Open-source code for local and hosted models

Keep in mind

The first edition leans on fast-moving tools, including the Assistants API and the author's own Nexus platform, and says little about performance at scale.

Key takeaways

  1. Agents are assemblies of distinct components, including personas, tools, memory, reasoning and planning, and each can be designed and improved separately.
  2. Multi-agent frameworks such as AutoGen and CrewAI let specialized agents divide work through conversation or defined roles and tasks.
  3. Behavior trees, a control structure from game AI, give agent workflows more predictable, inspectable execution than free-form conversation.
  4. Retrieval-augmented generation and explicit memory types give agents durable knowledge beyond a single conversation's context window.

Notable facts

About the author

Micheal Lanham

Micheal Lanham is a software and technology developer with more than 20 years of industry experience, described by Help Net Security as lead AI developer at Brilliant Harvest. He has written books on deep learning, game development and augmented reality, including Manning's Evolutionary Deep Learning and O'Reilly's Practical AI on the Google Cloud Platform.

Questions

What is AI Agents in Action about?
It is a hands-on Manning guide to building AI agents powered by large language models and connecting them into multi-agent systems. Micheal Lanham works through OpenAI's GPT Assistants, AutoGen, CrewAI, behavior trees, retrieval-augmented generation, memory, advanced prompting, and planning and feedback, with each chapter extending working Python code from companion repositories.
Should I buy the first or second edition of AI Agents in Action?
Most readers in 2026 should choose the second edition, published by Manning in June 2026, which rewrote more than 95 percent of the text and covers the Model Context Protocol, the OpenAI Agents SDK and agent evaluation. The 2025 first edition still teaches useful concepts, but several of its tools have changed substantially since publication.
Who should read AI Agents in Action?
Developers comfortable with Python, GitHub and LLM APIs will get the most from it, along with technical leads evaluating agent frameworks for their teams. It is not written for nontechnical readers: the book assumes familiarity with large language models and spends most of its pages on code and configuration rather than strategy or theory.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.