Playbook

The AI Product Builder's Playbook

AI didn't remove the process. It made it mandatory.

Start readingDownload the PDFPDF · 1.1 MB

A field guide for the new kind of builder: the person who owns the whole loop, from a vague idea to a product in production, without being a career developer. You make the decisions. AI does the execution. You keep it honest.

Everything here comes from real work — apps built for clients and for myself, from SaaS products to internal tools. None of the rules are abstract; each one was paid for by a real failure, on a real day, with a real cost. I keep the lessons and drop the names: the value is in the principle, not the address.

How to read this

The body is written to be read by someone who is not a developer — you don't need to code to understand any of it. Where a topic has a deeper edge, it's marked “For the CTO in you”, so you can go as deep as you want, or skip it. The parts run in the order you actually build: the role, the laws, the foundations, the work cycles, the proof, the shipping, the AI-specific stuff, the traps, and the glossary.

Contents

  1. 00The role: the AI Product Builder1 min read
  2. 01The laws of building5 min read
  3. 02The dev foundations (plain-language)3 min read
  4. 03CI/CD and mechanical checks2 min read
  5. 04The work cycles2 min read
  6. 05Specs and scoping (the core craft)3 min read
  7. 06Tests and QA: proof against intention2 min read
  8. 07Data, migrations, and money1 min read
  9. 08Shipping, security, and operations4 min read
  10. 09Documentation as infrastructure1 min read
  11. 10Design system, content, coherence1 min read
  12. 11Building with AI: what actually changes3 min read
  13. 12The catalogue of traps1 min read
  14. 13Going further (what I haven't done yet, and what dev+++ looks like)2 min read
  15. 14Glossary3 min read
  16. 15The self-assessment2 min read
  17. 16Case notes (anonymized)1 min read