AI Gave Me a Skill I Didn’t Have as a Designer

Because I design products, I knew the first step should not be “build me a QR platform.” It needed a clear brief first.

I wanted a free QR code generator that runs entirely in the browser, without a backend, without login, and without user data leaving the device.

I started with an MVP - the absolute basics, just to test whether the concept worked. No extra features, no polishing details, only the core of the product. Once the foundation proved itself, I started adding more layers.

First, a classic generator with the option to download the QR code as SVG or PNG. Then different types of QR codes - URL, vCard, WiFi, event, and others. Later, I added the ability to visually customise the QR codes.

Once the tool was working, I started expanding it further. I added language versions, FAQ, usage examples, ready-made templates, a blog, analytics, an AdSense experiment, landing pages, and eventually machine translation into 22 languages.

Each of these layers was a small experiment of its own. And each one was built in dialogue with AI.

This was not “write a prompt and you are done.” It was a process. A large part of the time, I was not only asking how to code something. I was asking how it could be improved, what could be added, where a problem might appear, and what possible solutions existed.

AI suggested options. I used my experience with digital products to choose what made sense, what was unnecessary, and what needed to be adjusted. And this is where it became clear that even though I do not know how to code, I still had an important role in the process. I had to define the task, break it down into smaller parts, check the result, decide on priorities, and not push everything to production just because AI labelled it as finished.

My role was not to write code.
My role was to lead the product.

And that is the most interesting part of the whole project for me. I did not need to learn programming first. I needed to know what I wanted to build, how the result should work, and how to verify whether it actually worked. AI helped me fill in the technical part.

It was not a smooth ride. And the friction was probably the most valuable part of the whole experience.

Translation was much harder than I expected. Machine-translating the content into 22 languages consumed a huge number of tokens. It was not the hardest part technically, but it was demanding in terms of volume. It was slow, expensive, and I had to split the whole process into smaller parts.

AI can quickly add more, but more does not always mean a better product. While expanding the project, more ideas, templates, pages, and options kept appearing. At first glance, it looked like growth, but not everything that can be generated automatically has real value. The decision of where to add and where to reduce had to stay with me.

“Done” did not always mean done. This was one of the most important lessons. Sometimes AI said a task was finished, but after checking it, I found that only part of the brief had been completed, or that it had chosen its own solution instead of following exactly what I asked for.

Several times, I directly asked whether all the points from the brief had really been completed. And the answer was essentially: no, some parts were not finished yet.

The reason was simple - it was saving tokens, shortening the process, or assuming that the proposed solution was good enough. That is why every larger task had to be checked point by point. Not trusting the word “done,” but verifying whether the result actually matched the brief.

Generated code does not mean a finished product. Without testing, clicking through the flows, and checking things repeatedly, mistakes could very easily have made it into production. AI can significantly speed up the work, but it does not remove the responsibility for the result.

Let’s be fair. It cannot really be monetised through AdSense. Google classified it as “low value content,” and in principle, I understand why. It is a simple tool, not a content project with deep added value.

And it is also fair to say that for most people, it would probably not be worth putting this much time into a similar project. It is not a breakthrough idea, and it probably will not become a business.

But I did not build it primarily for money. I built it out of curiosity. I wanted to know whether I could, as a designer who does not know how to code, build a functional product, put it online, and watch people from different countries use it.

And the answer is yes.

Today, I have my own tool that works, is available online, and is used by people around the world. Not because I learned how to code, but because I knew how to design a product and AI helped me overcome the technical barrier. For me, that is the real point.

Not QR codes.

But the fact that AI gives people new abilities. An idea that would previously have stayed only in my head because “I don’t know how to code” can now be tested, built, and released.

That barrier has become much smaller. And that changes who gets to create.

Try it live: qrcodeeasily.com