We tried to prove tenant isolation before tenants existed.
The running system exposed a hidden assumption and changed the architecture.
commercialization / work in progress
I work on commercialization.
How companies understand markets, find and create demand, position themselves, sell, learn and grow.
Today, I’m especially interested in how AI changes what is possible across that system.
on my desk
Questions I’m working on now.
testing
Can law firms detect demand earlier?
Signals, context, fit and learning.
↗exploringWhat happens to learning when people stop meeting?
Co-generated learning in hybrid organizations.
↗testingWhich AI model is adequate for which task?
Capability, cost, verification and workflow.
↗independent exploration
Can a travel company see demand before the search begins?
Not started yet. The question is on the desk.
recently changed
The work moves when the evidence moves.
Signal volume became less interesting than signal convergence.
The question shifted from counting events to testing whether independent observations support the same hypothesis.
Model cost changed how we structured the work.
More AI capacity did not automatically produce more useful progress.
2014
→Who matters — and what do they need?
2020
→What if marketing and sales are not separate problems?
2024
→What happens if the organization does not learn?
2026
What changes when AI can participate in the loop?