26 February 2026

Analysis of the book Game AI Pro 360 – Guide to Architecture

Book cover
There are many books about artificial intelligence (AI), but not so many focused on its use in video game development. Among those, there are two types: introductory books and in‑depth ones. The one we are dealing with belongs to the second category. That’s why I don’t recommend it to anyone who hasn’t already sharpened their teeth on a good number of introductory books and tried implementing AI algorithms. Believe me, this book is not for starting to learn AI. In fact, none of the books in this collection are. With that warning out of the way, I’ll try to justify it by explaining what the book is about.

The book does not follow an overarching narrative. It will not guide you through a learning path. Instead, it is a compendium of expert articles, each with a format and level of detail very similar to academic papers. If you’ve ever had to process the RFC of a standard, the articles in this book will strongly remind you of that. The different articles focus on different implementations of decision‑making systems (finite state machines, behavior trees, task planners, etc.), although each approaches them differently.

It’s clear that the authors of each article are true experts. No one can dispute that. Each of them refers to implementations they have made in well‑known games. The problem is that, for whatever reason, most of the articles leave me with the feeling that the authors are holding back their best cards—that they’re not telling you everything. Sure, I imagine that a full explanation of the complete AI implementation for those AAA games wouldn’t fit in a book, let alone in a single article, but I finished the book with the impression that unless you are as much of an expert as the authors, you won’t be able to fill in the gaps they leave in their explanations. It’s a constant feeling of “Okay, I think I understand at a high level what you want to do, but how exactly is that implemented?”—and that’s where the article ends. It’s true that sometimes you get code snippets, but they tend to be fragments of interfaces or high‑level pseudocode.

I suppose we should at least be grateful that this book is a topic‑based compilation of the articles previously published in the earlier Game AI Pro series. At least the compilation has some thematic coherence. I didn’t read the older books, but if everything was mixed together, they must have been very confusing.

For all these reasons, I must warn you against this book if you are just entering the world of AI for video games. If that’s your case, I don’t think you’ll get anything useful from it. There are better options in which to invest your money and from which you will learn much more. I would only recommend this book to someone who has already been working professionally in this field for many years; and even then, I’m not sure whether the book’s content hasn’t already been surpassed by time.