It started with a birthday present.
My wife had been talking about a venture she wanted to start, and I thought I'd surprise her with a book — something written entirely for her, covering the ground she'd need to take that first step. I wasn't thinking about anything fancy. I had the content, I had a PDF, and I just wanted to print it out and hand it to her. She loved it. Getting it printed, though, was a different story.
The birthday present
I took the file to a local print shop. The woman behind the counter looked at it and asked where the cover was. I didn't know I needed one. She suggested Canva, and suddenly what had been "print this out for me, please" turned into a design project. I opened Canva and I could immediately see the potential — beautiful templates, drag-and-drop tools, all very impressive. But I wasn't designing a poster. I was trying to wrangle a hundred-and-fifty-page book into something printable, and Canva fought me on every page. It was not built for this.
I looked for something better. Found a piece of software designed specifically for formatting fiction books. Paid for it. Spent an evening trying to make it work with my content. It wasn't right either. I returned it.
Then I tried the technical route. Quarto and Typst — coding tools that let you define layouts programmatically. AI helped me write the code, debug it, redesign it. I thought I was getting somewhere. Twenty-five hours later, I had a book that looked acceptable. Not good. Acceptable. For a birthday present that was supposed to be a nice surprise, I'd turned it into a part-time job.
The thought that changed everything
Somewhere around hour twenty, sitting in front of a screen full of code that was doing something wrong with the page margins — again — I had a thought that changed everything.
This can be automated.
Not just the formatting. The whole process. If I could describe what a well-designed book should look like, and AI could help me build the system to produce it, then I'd never have to fight with page margins again. Nobody would.
That thought became a project. The project became a platform. Two months later, I had a working publishing engine — software that takes a manuscript and produces a properly typeset, print-ready book. It has close to five hundred automated tests. It handles running headers, tables of contents, dynamic footnotes, all the fiddly things that had cost me those twenty-five miserable hours. This is the second book it has produced.
A note on background
I should be clear about something. I'm a quantitative risk modeller by trade — I've spent years in credit risk, building models in SAS and R, working with data professionally. So I'm not starting from zero with analytical thinking or even with code. But building a full software platform is a completely different category of work, and I could not have done it without AI. More importantly, you don't need my background to do what I did. The coding knowledge helped, but it wasn't the point. The point was knowing what I wanted, being able to describe it clearly, and being willing to keep pushing until it worked. Those are human skills, not technical ones.
That's what this book is about.
Not the publishing platform specifically, but the thing that made it possible. When people talk about AI, they usually talk about doing things faster — writing emails quicker, summarising documents in seconds, generating images on demand. And AI does all of that. But the interesting part, the part nobody told me about, is what happens after you've been working with it for a while. You stop thinking about tasks and start thinking about systems. You stop asking "how can AI help me do this?" and start asking "why am I doing it this way at all?"
That shift — from faster tasks to better systems — is what took me from a printing disaster to a working platform. It's also what helped me build an investment management system, navigate a career transition, and write a book about the whole experience. All of it required a willingness to start, to be specific about what I wanted, and to treat AI as a genuine thinking partner rather than a fancy search engine.
Who this book is for
You've used AI enough to know the problem: sometimes it's brilliant, sometimes it's useless, and most advice about it doesn't help much. You haven't found your way into using it properly, and most of what you've read about AI is either breathless hype or dense technical material that assumes you care about how the technology works.
You don't need to care about how it works. You need to know how to work with it.
This isn't a technical manual. There's no jargon in these chapters, no talk of models and parameters and context windows. If you want that, there's an appendix at the back. This is a book about thinking — how to approach AI as a collaborator, how to get genuinely useful results, and how to go from "that's interesting" to "I built that."
Everything in this book comes from things I actually did, mistakes I actually made, and conversations I actually had with AI tools over the past several months. I'm going to show you what worked, what didn't, and what I wish someone had told me at the start. You don't need to use all of it — any single idea in these chapters will improve how you work with AI, and each one you add makes the next one easier.
The only thing you need to begin is a problem worth solving and the willingness to have a conversation about it.
Let's start.