When the team splits
The fight about features is really about what the product is for.
Making things got cheap. Deciding what is worth making did not.
I help founders and small teams work out where AI changes their numbers, and where it doesn’t.
Right now I’m advising GrönLoop, a Swedish circular-economy startup.
I also teach and shape curriculum on the product design programme at the Estonian Academy of Arts.
Currently building Jon and BTO, and doing an Executive MBA with AI at Arden University, UK.
I’m most drawn to work around wellness, education, equality, kids, and deep-tech.
How can I help?
Most products don’t fail at the build. They fail at a decision made long before it. That decision is where I help.
If one of these is your situation, tell me which one.
Before the build
You’re about to spend big, and nobody knows when it pays back.
After it works
It works. Nobody can say why anyone would pay for it.
When there is nobody to build it
You need a working version, not another document.
I set up GrönLoop’s first MVP and ran the pilots that validated the basics. AI tools are starting to write reports for free, so we moved the product to the handover, the moment waste passes from one company to another and the data is first created. Next: an AI agent at each handover, so no company has to log in to someone else’s system.
- Product direction and positioning
- What to build now and what to leave manual
- A two-track plan for revenue and long-term position
- Revised targets and a pricing approach
- Which regulations and partners the product has to fit
BTO, Estonian Academy of Arts
Defence grading is usually a mess of paper sheets, private spreadsheets, and numbers reconciled after the fact. Evaluators can't see where the committee stands, aggregates are calculated by hand, and the reasoning behind a grade is gone the moment the session ends. BTO replaces that with a single live system, in use now and still being built out.
- Product definition and scope
- Interface and interaction design
- Built and shipped on my own
- Real-time grading for the whole committee
- In use at the Estonian Academy of Arts
Jon, a tool for learning demanding material
Jon turns your own material into daily practice. You write your concepts in your own words, then Jon hides the source and asks you to explain them back from memory. It is built on a 2013 study that ranked ten ways of studying by evidence. Rereading and highlighting came out lowest. Jon uses the ones that came out highest.
- Retrieval, teach-back and spaced repetition, the methods with the strongest evidence behind them
- The material stays on screen while you build. It is hidden while you practise
- Jon reads your concepts against the source and tells you what you missed, but never writes them for you
- In testing with four people, on their own material
Building something deliberately unglamorous. Not the work that looks impressive in a portfolio, but the work people actually need. It pulls together everything I already do, and it gives me the freedom to choose the rest.
Exploring what design education should be in the AI era. I build courses on the product design programme at the Estonian Academy of Arts and use them to test what holds. The research and the writing are where I work out which part of design survives the automation, and what has to change in how it is taught.
- Course design and curriculum
- Product design, portfolio and product management
- Admissions and thesis committees
- Mentoring
