Operating model
Built to learn fast.
I use AI as a core part of how I work in product, from strategy and discovery through design, build, and delivery.
The goal is not personal productivity. It is making it so fast and easy to test ideas that teams learn continuously instead of shipping on opinion.
The problem
Most product organizations are slow to learn, not slow to build.
- Research sits in one tool,
- design in another,
- requirements in a third,
and by the time an idea reaches a customer it has been through so many translations that what comes back is hard to read.
So teams ship on opinion and find out later.
A different way of working
One codebase
The design system, customer journeys, evidence, and product decisions live in one place, in code, and product, design, and research build in it together.
Out of documents
Collaboration moves out of documents and into the product itself. Anyone can open any part of it, use it, and add to it, long before it is built.
Days, not quarters
Ideas get in front of customers and stakeholders in days rather than quarters, so bets are validated or killed while they are still cheap to change.
Evidence over opinion
Ideas reach customers early enough that data, not seniority, settles the call.
Collaboration in the product
Teams and stakeholders explore and discuss the working product together, not documents about it.
Experience it before it exists
Anyone can open any scenario and use it, rather than read about it.
Validate before you commit
Test and learn before significant engineering investment.
The hardest product decisions are rarely about what to build. They are about identifying the right problems and defining success before committing to delivery.
Where the craft is heading
Product managers, designers, and researchers are no longer specifying work for someone else to build. They are building, testing, and learning directly, and the distance between understanding a customer problem and putting something in front of them keeps shrinking. That is where the craft is heading, and it is how I build the teams I lead.
Principle
AI makes learning fast. What to build, and what to kill, stays human.
Tools
- Claude
- Lovable
- Codex
- Figma
- GitHub
- Vercel