AI Agents in Action versus Building Agentic AI Systems.
Both show up on every "best" list. They're not competitors. They're a sequence. Here's which one to read first, and when.
Reviewed by Ashish Sheth · Updated April 2026
Author
Micheal Lanham
Anjanava Biswas, Wrick Talukdar
Pages
344
286
Published
2025
2025
Publisher
Manning Publications
Packt Publishing
Level
intermediate
intermediate
Amazon Rating
4.1/5 (35)
4/5 (69)
Goodreads Rating
3.11/5 (74)
2.93/5 (27)
AI Agents in Action
Strengths
+ Practical, code-first approach with runnable examples throughout
+ Manning's quality editing makes it more readable than rushed competitors
+ Covers OpenAI Assistants, AutoGen, and LangChain in one volume
+ Author's game-AI background brings useful patterns (behavior trees) to LLM agents
Caveats
− Some readers find the OpenAI Assistants focus dating quickly
− Coverage of newer agent frameworks (LangGraph, CrewAI) is thinner than expected
− Code examples occasionally hit API breakage as providers evolve
Building Agentic AI Systems
Strengths
+ First-mover book on a rapidly emerging topic with few good resources
+ Authors bring real production experience from Amazon and AWS
+ Covers practical patterns (CWD, reflection) that map to LangGraph and similar frameworks
+ Strong on safety and ethics, often overlooked in agent literature
Caveats
− Some readers report repetitive and over-explanatory writing
− Code examples are uneven, some feel rushed
− Field moves so fast that some specific framework references already need updating
The verdict
AI Agents in Action is the stronger pick overall, but Building Agentic AI Systems may suit you better if you're a AI engineers building agent systems.
AI Agents in Action
Check Price on Amazon →
Building Agentic AI Systems
Check Price on Amazon →
Frequently asked
Which is better, AI Agents in Action or Building Agentic AI Systems?
AI Agents in Action is the stronger pick overall, but Building Agentic AI Systems may suit you better if you're a AI engineers building agent systems.
How is AI Agents in Action different from Building Agentic AI Systems?
Lanham's book is more code-first and beginner-friendly. Biswas/Talukdar's book is more architectural and patterns-focused. Read Lanham to build, Biswas to design.
Do I need to know LangChain or LlamaIndex first?
Familiarity helps. The patterns transfer to LangGraph, AutoGen, and Bedrock Agents, but the book assumes you know what an LLM call and a tool definition look like.