How I work
From a loose idea to a product people can test and investors can see.
I work with founders and small teams who have ambition and not enough structure yet. This is the full method: how I find the problem worth solving, cut it down to an MVP, and prototype it well enough to test with users and pitch to investors.
What I design against
A good MVP sits where three things overlap.
What people struggle with.
User evidence first. Interviews, support logs, analytics, reviews, and the steps where people give up. I look for pain that is frequent, costly, and already being worked around.
What is now possible.
AI models have opened product territory that was too expensive or too slow to build two years ago. Much of it is still unexplored. I check which new capabilities solve a real pain, and which ones are just a demo.
What the market will back.
A problem is only worth building for if someone will pay and the timing is right. I look at who is already solving it, how the current funding climate treats the category, and what a small team can credibly ship.
Most products fail on one of these. The method below checks all three at every stage.
The method
Open the problem up, narrow it down. Then do the same for the solution.
Each stage ends with something concrete you can use, share, or decide on. You can stop after any stage and keep what we made.
Stage 01. Map the whole journey
I start wide. Every step from first hearing about the product to coming back a second time, including the parts outside the screen. Founder assumptions go on the map too, marked as assumptions.
Where AI helps
Synthesising interview transcripts, reviews, and support tickets into themes, so more of my time goes to talking with people.
You get
A journey map with pain points ranked by how often they happen and how much they cost the user.
Stage 02. Name the pain worth solving
Then I narrow. We pick the one to three pain points that matter most and write down what we are deliberately not solving yet. This is where most scope creep gets stopped.
Where AI helps
Stress-testing the problem statement against the evidence, and surfacing edge cases we would otherwise meet in development.
You get
A problem statement, the evidence behind it, and a written list of what is out of scope.
Stage 03. Explore more than one answer
I open up again, this time on solutions. I develop a few directions, including ones built on new AI capabilities, and check each one against user value, build effort, and market case. Engineers join here, not after.
Where AI helps
Fast rough variations and quick technical spikes to see what a model can actually do with real inputs.
You get
Two or three directions compared side by side, with a recommendation and the reasoning.
Stage 04. Cut to the MVP and prototype it
I narrow to the smallest feature set that proves the core promise. Then I build a clickable prototype with realistic content, empty states, and errors, detailed enough that users react to it as a product and investors can see it working.
Where AI helps
Front-end prototypes and variant flows built in days, so we test the real interaction instead of static screens.
You get
A prioritised MVP feature list with the reason for each feature, and a working prototype ready for testing and pitching.
Stage 05. Test, learn, redesign
I put the prototype in front of real users, sort the feedback into what changes the product and what is noise, and redesign. Then we run the loop again. Each round is documented, so you can show investors how the product earned its shape.
Where AI helps
Clustering feedback across sessions and tracking which issues repeat between rounds.
You get
A change log of what we tested, what we learned, and what changed as a result.
Stages 03 to 05 repeat until the prototype holds up with users.
For founders raising
A prototype that answers the questions investors ask.
Investors are not buying screens. They want to see a real pain, evidence that people have it, a reason this is possible now, and a team that knows what not to build. The method produces each of these as a by-product, so the pitch prototype is the design work, not a separate deck exercise.
- Is the problem real?
- Ranked pain points and the evidence behind them (stages 01 and 02).
- Why now?
- The technology and market check in stage 03.
- Can this team ship it?
- A tight MVP scope with a written out-of-scope list (stage 04).
- Do users want it?
- Test rounds and the change log (stage 05).
- Can I see it?
- A working prototype, not a slideshow.
Why I understand this side of the table
I have been the founder, not only the designer.
I am the founding designer at OHR.FM, a music culture platform, where I work directly with the CEO and CTO on product direction. Before that I founded Altrue, an AI-guided microdonation product, and took it to a working MVP on my own.
I paused Altrue when it became clear that payments, fund distribution, and trust in a charity-adjacent space needed more than one person could responsibly deliver. That decision taught me as much as building it did. I now help founders see those walls early, before they spend a year running into them.
Read the Altrue case study →Before freelancing: 9+ years at Airalo, IKEA, H&M, and Husqvarna, through Frog Design and in-house.
AI in the method
AI shortens the distance to evidence. The decisions stay mine.
Prototype earlier.
Realistic, clickable flows before a team commits to build.
Stress-test product logic.
Edge cases, empty states, onboarding paths, and variant flows, found before development instead of during it.
Edit with judgement.
AI produces more options. My job is to narrow them, protect the user, and decide what not to ship.
Engagements
Start with one stage or run the whole method.
Problem sprint
Stages 01 to 02
For founders with an idea and too many possible directions.
Duration: 1 to 2.5 weeks
MVP and pitch prototype
Stages 01 to 05
For founders preparing to raise or to start building.
Duration: 4 to 8 weeks
Redesign loop
Stages 03 to 05 on an existing product
For products that are live but not landing.
Duration: 3 to 6 weeks
Workshop
A single stage, run with your team
For teams that need alignment in one or two days.
Duration: 1 to 2 days
Have something half-formed? That is the right time.
Send rough notes, a deck, or a link. Tell me what you are trying to validate, build, or pitch. I reply within 2 working days.