You Can Build That

Maybe it's a form your customers keep emailing about, a sign-up sheet that still lives on paper, or a spreadsheet you retype every week. AI tools can now write the code for small apps, websites, and automations from plain-English requests, and you don't need to learn to program to use them.

But AI work is often almost right, and almost right is where the trouble starts. You Can Build That shows beginners, not developers, how to build small, useful tools carefully, using one simple method, SKETCH: Start small, Know what done looks like, Edit one step at a time, Test like a user, Control the keys, and Hand it over with care.

In this book, you'll discover:

• How to choose a first project you'll actually finish
• How to write a one-page spec and acceptance checks so the AI builds what you meant
• How to work with AI one step at a time, with saved versions you can go back to
• How to test like a real user and read error messages without panic
• How to keep API keys, logins, and personal data safe
• How to put a tool online and keep its costs and limits under control
• How to maintain what you build, and when to bring in a professional developer

Build small, check everything, and make something genuinely useful.

Why read this book

• It is written for people who have never written code: plain language, no programming lessons, and no hype.
• It is tool-agnostic, built on habits that keep working as AI chat assistants, app builders, coding agents, and automation tools change.
• It is grounded in published research and documented incidents, and it is honest about what AI gets wrong, from security holes to answers that are almost right.
• Every section ends with one small "Try this" step and sample requests you can type into an AI tool.

AI disclosure: The text of this book was generated with AI (Claude by Anthropic), and the cover image was created with AI.