Some ideas do not fail. They just wait. Mine waited about six years, and what finally released it was not a better plan or more free evenings - it was AI.
It started with a poster
Herning. A fintech company. I was leading a Tech Academy, which meant my job was to make learning something the organisation actually did rather than something it put on a slide once a year.
One of the teams had a poster on the wall. A grid: names down one side, skills across the top, marks in the cells. A skills and systems matrix. Nothing sophisticated - printed, taped up, updated by hand.
But it did something I had not seen a single tool do. It made knowledge visible. Standing in front of it, you could see in five seconds that only one person in the team knew a particular payment system. You could see who wanted to learn what. You could see where the team was thin - and crucially, the whole team could see it at the same time. It was not a manager reading a spreadsheet; it was a shared picture that people argued about, pointed at, and planned to improve.
That poster stayed with me long after I left.
Ballerup, React, and a problem I gave myself
A year later I was at an insurance company in Ballerup. The frontend was built in React. I did not know React.
I could have read the documentation and worked through a tutorial. Instead I gave myself a real problem to solve, because that is the only way I have ever properly learned a technology: build something you actually want to exist and fuel the learning with a goal.
So I built the poster. The first version of a competency matrix as software - skills, people, levels, gaps. It taught me React, and it turned a wall-sized artefact into something a distributed team could keep up to date.
The job offer I turned down
I showed it to an AWS consultancy. The conversation ended with a job offer.
I said no. It was not the right move for me at that point. But the offer told me something useful: the idea held up outside my own head. Somebody who had no reason to be polite about it looked at the thing and wanted the person who built it.
Then I put it in a drawer.
Several years in the drawer
Not abandoned - parked and occasionally dusted. This is the unremarkable economics of a side project: a full-time job, a family, and a gap between knowing what to build and having the hours to build it. The idea and what to build was never the bottleneck. Execution time was.
What changed
AI changed it. Specifically, building with Claude Code.
The honest version is not "AI wrote my product." The honest version is that the distance between an idea and working software collapsed. Work that used to consume three evenings now took three hours. The scaffolding, the boilerplate, the parts of the job that were never the interesting bit - they stopped being the reason not to start.
A side project only survives if progress is visible between one evening and the next. Suddenly it was.
The drawer opened.
Komforan as a side gig
I founded Komforan ApS as the serious side project, and as plan B, and launched www.komforan.io - the competency matrix, finally live, finally real. In the process I had fun, learned a lot, and figured out that solving problems and making the solutions come alive is what I enjoy the most. So Komforan built MockyMe and Tolkra as well, and the original komforan.io website turned into Kunska.
Kunska
So now Komforan builds several products, with more coming, and none of them should be hiding behind the company name. The competency matrix has a place of its own.
www.komforan.io became www.kunska.com - one product in the Komforan portfolio, with its own name and its own identity.
Same idea as the poster in Herning. Make what a team knows visible, so the team can do something about it.
What I take from it
- Good ideas do not expire. They wait for a constraint to lift. Mine waited for the cost of building to fall.
- Learn by building something you want to exist. React went in far deeper through a project I cared about than it ever would have through a course.
- Pay attention to the low-tech thing that works. A paper poster beat every tool I had seen. That was the signal.
- "No" is a legitimate answer to a good offer - as long as you keep what the offer taught you.