How to Use This Book
This book is written for executives, managers, and EMBA students who need to do more than talk fluently about generative AI. My ambition for you is specific, and it is worth stating on page one: by the end of this book, you should be able to work the way a Forward Deployed Engineer works. That means you can walk into a business unit, find the workflow where AI will actually pay, define what "good" means in measurable terms, get a working prototype in front of real users within weeks, prove or disprove its value with evidence, and govern what you deploy. Chapter 14 describes that role and method in full; every chapter before it builds one of the capabilities it requires.
The book has three parts. Part I (Chapters 1–3) covers foundations: what the technology is, what the field evidence says about adoption and value, and the five engineering disciplines for working with models, from prompting to loops and graphs. Part II (Chapters 4–7) is the builder's toolkit: the model landscape, API integration, the specialized tool ecosystem, and the Five A's framework for choosing the right level of automation. Part III (Chapters 8–14) is strategy and execution: business functions, the EDGE value framework, the implementation roadmap, business model innovation, what is coming next, the ethical and regulatory landscape, and finally the Forward Deployed method that ties it all together. An appendix supplies the working artifacts, a pilot charter, an evaluation design, an ROI model, and a governance intake form, that you can copy and use in your own organization this quarter.
Three habits will make the book more useful. First, each chapter opens with learning objectives and closes with discussion questions; if you are reading with a study group or a class, argue about the questions, because the arguments are where the learning is. Second, the field moves quickly, so specific model names and prices in this book are date-stamped snapshots (mid-2026); the durable content is the frameworks, the selection criteria, and the habit of checking current sources, which Chapter 4 teaches. Third, do not just read: pick one workflow in your own organization and run the six-week playbook in Chapter 14 as you go. A reader who finishes this book with one deployed (or honestly killed) pilot has gotten far more from it than one who finishes with highlights.
A note on evidence and disclosure. Where this book cites numbers, it names the source and the year of fieldwork; where a claim is my own judgment from consulting and teaching, the first-person voice makes that clear. Some examples in the tools chapters (GAIforResearch, Mimi, Catlendar) are projects I built; they are labeled as such where they appear, and they are there because I can speak to their construction honestly, not because they are the best in class.