What this is
AI is not another tool we cram into the model. It’s a force that changes what product management is. Which decisions we make. How teams are built. What separates a good product organization from a bad one.
This playbook is a framework for that shift. Five layers, from culture down to tools. Each with a principle, and something I’m still not sure about.
Written for product managers, designers, engineering managers and heads of product. Read it front to back, or jump straight to the layer that’s pressing on you most.
Who it’s for
Product managers, design leads and engineering managers who are responsible for how their team actually works. Not people who want to learn to prompt better. People who wonder whether the whole frame around their team still holds.
It also speaks to CTOs, VPs and founders who need to rethink structure, roles and processes from the ground up, because the old model has stopped making sense.
If you’re looking for an “AI transformation” checklist, this is not for you. If you’re looking for an honest conversation about what actually needs to move, and in what order, read on.
Why this pyramid
My thesis is that most AI transformations start at the bottom, with tools. “We rolled out Copilot.” “We got Cursor licenses.” “We use ChatGPT Enterprise.”
That’s the fastest route to a failed transformation.
When tools land in a culture that hasn’t moved, an organization still built around coordinating execution, and roles defined by what people produce, everything that creaks gets amplified.
Culture at the top, tools at the bottom. Each layer governs the one below it. If the culture isn’t in place, reorganizing won’t help. If the organization isn’t in place, learning new skills won’t help. And so on.
Five layers:
- Culture & identity. Mental models, beliefs, who we are
- Organization & structure. Teams, roles, collaboration
- Skills & practice. What people know and do
- Processes & workflows. How the work flows
- Tools & software. Agents, code, platforms
It helps you sort your thinking and figure out where something actually needs to move.
Four things this whitepaper stands on
My own practice. I ship products with AI every day: Stilnote, Vibetjek, Knud, Dynamo, Super Yes. This is not theory. It’s reflections from a desk where Claude Code runs all day.
Consulting work. I’ve seen how organizations like Norlys, Energinet, LEGO and Jyske Bank try to navigate the product transformation. Where it works, and where it goes wrong.
The product management literature. Cagan, Perri, Cutler, Skelton/Pais, Microsoft Research and others. I try to connect to the voices that have already said something true about the field, rather than pretending everything is new.
My own history. I’ve been writing about product roles and UX since 2022.
What you get out of it
No “top 10 AI tools.” No one-size-fits-all process.
Instead:
- A framework you can use to sort your own organization’s challenges
- A vocabulary for conversations with leadership and colleagues
- Concrete principles you can test against your own reality
- An honest reflection from someone who builds and advises in the field
Format
Five layers, each around 1000 words. Read it as one whitepaper or as standalone articles. Each part ends with a principle, something that can stand on its own.