Building Agentic AI Systems is one of the clearest books I’ve read on the future of AI. The authors do a great job of explaining how autonomous agents can reason, plan, and act in real-world settings without drowning you in jargon. I especially appreciated how it balances technical concepts with practical examples in areas like automation, finance, and healthcare.

What really stood out was the focus on trust, safety, and responsible AI, which makes it feel both relevant and forward-looking. It’s not just theory — the examples and frameworks make it actionable for anyone curious about how to build or apply agentic AI in practice.

Highly recommend if you want to understand not just the hype around AI, but the real way it’s being built and used.

Good book, for beginners, nothing special or advanced

Authors have crafted Building Agentic AI Systems into a rare blend. Grounded, code-informed, architecture-smarter, and ethically considered. If you’re looking to step beyond prompts and into building AI agents that reason, plan, and learn with safety and structure…this book isn’t just a guide, it’s a gateway.

This book is useless. It randomly picks a bunch of AI jargon and then defines them or provides examples of utility. Learnt pretty much nothing in the first 2 chapters and then stopped reading it at that point.

Great book for anyone wanting a structured overview.

Building Agentic AI Systems is an in-depth and forward-looking guide that expertly blends foundational AI theory with the design and implementation of intelligent, self-directed agents. It offers a structured journey from the basics of generative models like VAEs, GANs, and transformers to the sophisticated behaviors of autonomous agents capable of reasoning, learning, planning, and introspection. What truly sets this book apart is its hands-on practical focus. Through well-crafted examples—such as intelligent customer support bots, financial forecasting agents, and e-commerce price strategists—it provides concrete, real-world scenarios that practitioners can directly apply. The Coordinator-Worker-Delegator (CWD) model offers a clear, modular approach to building multi-agent systems with defined roles and collaborative workflows.

Beyond implementation, the book gives thoughtful attention to trust, safety, and ethics—discussing explainability, bias mitigation, privacy, and responsible design practices. Topics like agent memory architecture, state modeling, and contextual awareness are covered with both technical depth and clarity.

Whether you are an AI engineer, system architect, researcher, or developer, this book serves as both a practical playbook and a visionary roadmap for building responsible and scalable agentic AI systems. It’s a must-read for anyone aiming to shape the future of autonomous, human-aligned artificial intelligence.

I didn't like this book. It was too slow. It spent the first half of the book on theoreticals and high-level concepts, like it was written for execs or non-technical people. There may have been some great hands-on stuff in the second half, but I couldn't stick around to find out.

TL;DR: "Building Agentic AI Systems" provides a comprehensive roadmap for developing autonomous AI agents that can reason, plan, and adapt. This book bridges theory and practice brilliantly, making complex concepts accessible without sacrificing depth.

Aha! Moments:

The chapter on the Coordinator-Worker-Delegator model was a revelation for me. This framework for organizing multiple agents with distinct roles completely changed how I approach multi-agent system design. Instead of building monolithic agents, I now see the value in specialized roles working in concert.

Another eye-opener was the distinction between mere "tool use" and true "planning" in agent systems. The authors clarify how agents that just call tools are fundamentally different from ones that can strategize and sequence actions toward goals - a distinction I now see everywhere in current AI systems.

Standout Ideas:

The book's approach to "reflection and introspection" in agents is particularly innovative. Rather than treating agents as static systems, the authors demonstrate how they can analyze their own reasoning, learn from experiences, and improve decision-making - making systems that grow smarter over time rather than remaining fixed.

I was also impressed by the practical implementation examples using frameworks like LangGraph, CrewAI, and AutoGen. These concrete demonstrations show how theoretical concepts translate into working code, complete with comparison of different implementation approaches.

Usefulness:

If you're building AI systems that need to move beyond simple prompt-response patterns into truly autonomous behaviors, this book provides both the conceptual foundation and practical techniques needed. It's given me a structured way to think about agent memory architectures, planning algorithms, and multi-agent collaboration that I'm already applying in my work.

I highly recommend this book to AI engineers, AI product managers, and researchers looking to understand where AI is heading as it moves beyond foundation models to truly agentic systems.

Great !

In times when true advise is needed to cut through the hype - this book is extremely either extremely shallow or overstating dramatically ( rather than a grounded description of what LLM-based agents can reliably do today)

E.G. - if you believe this is the current state of technology: "A reflective travel agent can introspect on the potential impact of its recommendations and adapt its strategies to align with ethical norms and human values. (p.183)." go ahead - waste your money.

I am trying to return it for a few days now - Amazon for what ever reason doesn't let me.

Bom conteúdo.+

This book contains Solid foundational concepts for agentic AI . A good buy !

Bloated with fluff, very thin on ideas, and not SOTA.