Building · 4 August 2026 · 6 min
I'm a UX consultant. I shipped my own product.
Why I built Wonfolio, what changed in how I work with clients, and the one rule I would not break.
For ten years my job ended at handoff. I designed the thing, specified it, stayed close to development, and then someone else owned whether it worked.
Wonfolio is the first product where that is not true. I designed it, built it, shipped it, and I run it. Here is what prompted that and what it changed.
The problem I kept seeing
I coach designers. The pattern never varies: the work is strong, the write-up never happens.
Enterprise designers have it worst. Their best projects are under NDA, have no shareable screenshots, and require three paragraphs of context before the interesting decision makes any sense. Writing that up properly takes a weekend they do not have.
So the designer who solved a genuinely hard problem (a clinical data platform, a safety-critical HMI) loses the role to someone with a beautifully written case study about a food delivery app.
That is not a talent problem. It is a documentation problem. Documentation problems are solvable.
Wonfolio
You give it the raw material: a portfolio PDF, a CV, a folder of project notes. It gives you back a full case study of 700 to 1,000 words, a portfolio summary of about 150, and three LinkedIn posts. Roughly 10 minutes.
Free covers one case study, one testimonial request and your own portfolio page, with no card and no expiry. €12 a month lifts the limits, drops the branding, and adds application pages tailored to a specific job posting.
The rule I would not break
Wonfolio invents nothing.
No fabricated metrics. No invented quotes. No “increased engagement by 40%” that nobody measured. If a number is missing, the output says so rather than filling the gap.
This was the hardest product decision, and it was not close. An AI writing tool that invents plausible figures is genuinely useful right up until the first interview question, and then it is worse than useless, because the person has now lied on the record without meaning to.
The temptation is real. Invented numbers make the output look better in every demo. Refusing them means a visible gap where a metric should be. I took the gap.
You will find the same rule on this website. There are placeholders in my own case studies where an outcome metric belongs, because I do not have permission to publish the real figures. An empty bracket is more honest than a confident number nobody measured.
What actually changed in my client work
Three things, and none of them are what I expected.
I stopped being able to hide behind a spec. As a consultant you hand over a specification and the difficulty of implementing it becomes someone else’s problem. When you build it yourself, every “just” in your own documentation gets audited by reality. I write cheaper specifications now.
I understand the cost side of a decision much better. I always knew that a design choice had an implementation cost. I did not feel it. There is a category of elegant idea I now discard in about 30 seconds because I know exactly what it would take to maintain.
I can talk to engineering teams differently. Not because I have become an engineer; I have not. But the conversation changes when you have shipped something and operated it. You stop trading opinions about feasibility and start trading experience.
Should every designer build a product?
No. This is not a “designers should learn to code” argument.
But there is a specific version of it that I do believe: if you are going to consult on AI products, build one. The current market is full of people describing AI-product design who have never had to decide what their own system does when the model is uncertain, never watched their own error rate, never had a user hit an edge case they did not anticipate and email them about it.
Those decisions are the actual job. You cannot get them from a case study.
What it cost
Time, mostly. Evenings and gaps between client projects, over several months.
The strange part is that it made me better at the work that pays. I brought sharper judgment to the Siemens and Nordex copilot projects than I would have a year earlier, because I had been on the other side of the same decisions with my own product and my own users.
That is the return. Not the subscription revenue: the judgment.