I wrote another book. This book is a little different than my last one though.

It is called Do Agents Dream of Electric Flows? The Reality of Agentforce.

The title is a riff on Philip K. Dick’s “Do Androids Dream of Electric Sheep?” – the book that Blade Runner was based on. (If you don’t know either of those, I highly suggest checking them out.) And it fits – we are trying to figure out where the human ends and the machine logic begins in the Agentic era.

Why did I write this book?

There is a massive gap right now between what you see on a keynote stage and what you see when you open Salesforce Setup. Agentic AI and Agentforce are the current obsessions, touting unbelievable results with the wave of your magic AI wand. But when you actually sit down to build an agent, the magic fades and it immediately becomes architecture.

That is what the book is about.

A lot of the industry is treating AI like the solution to all your problems while also greatly reducing your overhead costs. It isn’t. If you plug Agentforce in without proper guidelines and constraints, especially without human oversight – it will be a catastrophe. The LLM handles the reasoning, but you still have to give it the foundation to ground it and the human to direct it.

I cover how this actually works. We look at the difference between deterministic execution and probabilistic reasoning. In a standard Flow, if A happens, then B happens. With an Agent, if A happens, the model decides if it should do B, C, or ask a question. That requires a totally different way of thinking and planning.

Let’s talk about the actual build process. You go in to set up a copilot or a custom agent. The first thing you realize is that prompt engineering isn’t just typing nicely. It’s systemic. You are writing instructions that have to account for unpredictable user inputs. It can be frustrating at first. You will test an agent, ask it a simple question, and watch it confidently execute the wrong action because your flow was exposed without proper parameters.

I didn’t write this to tell you AI is going to solve all your problems. Honestly, it might create new ones (potentially worse ones) if you don’t architect it right. I wrote this book to provide a baseline. The technology is moving fast. What Salesforce announces at Dreamforce or Connections changes the landscape every few months. But the underlying mechanics, like how an LLM interfaces with deterministic code, remain constant.

This isn’t meant to make you go “Wow, that AI sure is awesome!” The book is meant to make you feel empowered to decide for yourself where and when and how to adopt and implement AI in your org and your life.

So, What’s Actually In It?

Twelve Chapters. Dozens of movie and pop culture metaphors. And tons of information, opinions, and insights. No fluff.

I built my book around real archetypes and real groupings I have seen in the industry over my years inside of it. What seat you sit in drastically changes how you approach Agentforce and most books do not address this. Academic books give great ideas and theories, but in real life – a practitioner needs more.

I also have real war stories in here as well to help provide real life examples to give some color and context to the more academic information. I am not your teacher. I am not lecturing you from a classroom or a keynote stage. I am just a veteran practitioner sharing their experience with coworkers and friends at the local pub.

You can grab it on Amazon here: https://a.co/d/03KcDTHQ

NOTE: If you look at my book on Amazon prior to 9/7/2026 then it is only pre-order for the Kindle version. I will also have a paperback – but that is not able to be pre-ordered so it will not be available until release date.