Hey Reader,

In the first episode of my Mobbin mini-series, I talked about why AI-generated UI can end up looking generic: it's often working without enough real product context.

For the second episode, I put that idea into practice.

​

I redesigned a screen from Interview Prep Pal, a product I'm building to help designers practice for job interviews

The screen gives feedback after someone answers an interview question — what worked, what to sharpen, and what to try next.

It was doing the job, but the design felt flat. So I used Mobbin MCP to look at how real products handle feedback, suggestions, and critique, then brought those references into Claude Code to explore different directions.

I walk through the full workflow in my latest video.

​Watch episode two here.​

Interview Prep Pal isn't public yet, but you can join the waitlist here.​

Big thanks to the folks at Mobbin for sponsoring this video and supporting the mini-series.

If you've been curious about trying Mobbin, they're offering 20% off the annual Pro plan as a first-time subscriber.

​Try Mobbin here.​