Background
I often rely on navigation apps like Google Maps and Apple Maps. While great for directions, I’ve noticed a recurring gap: they don’t fully cover the needs of drivers looking for intuitive, reliable parking options. This is a common frustration I’ve shared with others in my area.
The idea
This concept explores how shared knowledge and small incentives can improve the everyday experience of finding parking.
Users contribute by sharing tips, flagging unclear restrictions, and rating locations – building a more trustworthy system over time.
To encourage engagement, a simple loyalty model rewards helpful contributions, creating value for both individuals and the community.
Process
When encountering something new, I tend to explore flows with different ways of generating ideas and sketching – but in this case I decided to explore how I could do this with AI, as much as possible.
Initially I tried out doing sketches with GPT, as you can imagine, it was pretty.. bad. It led me to believe it could create awesome Figma sketches, but that wasn't the case at all. It couldn't create images that looked anywhere close to what we had discussed either.
Then I tried Figma AI, wireframing drafts, and it was pretty bad also. But I'm sure it would have worked way better if I didn't already have a clear vision of what I wanted to achieve.
Therefor I just did the wireframing myself, with precision and speed. But…
…why did I choose mid-fidelity?
When things are complex or unclear, I usually start with low-fidelity to focus on flow and structure.
For this app, I had a clear idea from the start, so I moved straight into mid-fidelity and built a quick prototype in Figma.
Exploring tools like Figma Make, Lovable, and V0 later inspired me to push the concept further – creating a more immersive prototype to test with real users.
AI prototype
Prototyping with tools like Figma Make or Lovable can be incredibly powerful – though not always easy. In this case, it took some trial and error, but the end result was far more interactive than a standard Figma prototype.
My background in HTML, CSS, and JS was definitely helpful – but it’s still impressive how much can be achieved just by prompting.
I started by uploading wireframe screenshots to ChatGPT and asking for a Figma Make prompt. The first draft was decent, but I iterated further: specifying WCAG AA+ compliance, UX best practices, and adding extra screens and ideas. I also left room for suggestions – which opened up the concept even more.
Part of the results when asking ChatGPT for suggested flows and screens.
Once I had my carefully crafted prompt ready, I dropped it into Figma Make with high hopes. What I got? A single onboarding screen. That’s it.
I tweaked, reworded, and tried again – nothing. Switched to Lovable. Same story, different UI. Turns out, starting from pre-designed screens confused the tools more than it helped.
So I scrapped everything and started fresh, just describing what I needed instead. Ran out of credits in Lovable (of course), so I moved back to Figma Make and began working in small, focused steps – screenshot by screenshot.
It’s still a work in progress, but now I’m getting somewhere. Definitely not the journey I expected, but a better one in the end.
Screenshot of the desktop version of the app in Figma Make.
