Two ways to make your AI work the way your favourite book works

Most of us have that one book we keep going back to. It probably has a set of laws, a worksheet, or a way of thinking that we follow for a given task.

The problem is that none of that carries over into the AI. Every new conversation starts from zero, and you're back to explaining from scratch why you want things done a certain way.

If you want your AI to work the way the book works, there are two things you can do.

1. Turn the whole book into a skill

A skill is a structured file that your AI reads before it responds, so it has the book's knowledge as context instead of you having to explain it every time.

There's an open-source tool called book-to-skill (github.com/virgiliojr94/book-to-skill) that takes a PDF, EPUB, or any document and extracts the principles, mental models, and techniques into a skill file your AI can load. It pulls from across the entire book, so the AI gets the full picture, not just the chapter you remembered to paste in.

This works well when you want your AI to think the way the book thinks across different situations.

2. Extract one specific workflow from the book and make that into its own skill

Say your book has a diagnostic checklist or a decision matrix you keep coming back to. You take that one process, write out the steps the book uses into a file your AI reads before it starts the task, and now your AI follows that exact sequence every time.

You're not giving it the whole book, you're giving it one thing to do properly.

This one does one thing, and you can keep improving it every time you use it because you see where the steps are too loose or too rigid.

The difference

The first gives you breadth. The second gives you depth on one thing, and it gets better with use.

Both are a good way of taking knowledge you trust from a book and giving your AI the exact context to use it, instead of pasting paragraphs into a prompt and hoping it figures out what you meant.