How I Improved My Design Sense Using These 5 Claude Skills

I can't draw. Never could. And for years it showed in everything I shipped — my apps looked almost right in a way I could feel but never fix. So when people say you can improve design skills with AI, my honest first reaction was cynicism: I didn't want a robot to make art for me, because then the next screen would look wrong again and I'd be right back where I started. What actually changed my design sense wasn't AI making things. It was AI teaching me to see.

Here's the part nobody told me: taste isn't a gift you're born with. It's pattern recognition you build by having good and bad examples pointed out to you, over and over, until you start noticing them yourself. That's exactly the thing a handful of Claude skills turned out to be good at.

Can AI actually improve your design sense?

Yes — but not the way you'd guess. AI is genuinely bad at inventing a beautiful, original concept from nothing. What it's shockingly good at is critique: looking at what you made and telling you, specifically, why it feels off. "Your headline and your body text are almost the same size, so nothing leads the eye." "This spacing is inconsistent — 12px here, 20px there, no rhythm." That kind of feedback used to cost a design mentor and a lot of ego.

The shift for me was treating Claude not as a generator but as a tireless critic that never gets bored explaining the same principle a fifth time. Every critique was a rep. And after enough reps, I started catching the problems before I asked — which is the whole definition of having a better eye.

The 5 skills that trained my eye

These are the five that actually moved the needle, in the order that helped most.

1. Visual hierarchy critique. I'd paste a screen and ask, "What does the eye land on first, second, third — and is that the order I want?" Nine times out of ten the answer was no. My CTAs competed with my headlines; my headlines competed with a decorative icon. Learning to rank elements — one clear hero, everything else deliberately quieter — was the single biggest jump.

2. Spotting the generic-AI tells. This one's ironic and it's my favorite. I'd ask Claude to audit a design specifically for the clichés that scream "made by AI or a template": the purple-to- blue gradient on everything, three glassmorphism cards in a row, centered everything, a hero with a vague abstract blob. Naming those tells out loud made me allergic to them. Now I spot them in my own drafts instantly.

3. Spacing and rhythm. I never understood why designers obsess over an 8-point grid until Claude walked me through why inconsistent gaps make a layout feel amateur — your eye can't find a pattern, so nothing feels intentional. Asking "is my spacing on a consistent scale?" retrained me to think in rhythm instead of nudging pixels until it "looked fine."

4. Typography judgment. Type was my worst area. Too many weights, line-heights that suffocated the text, font pairings that fought each other. A skill focused on type would flag "you're using four weights where two would read cleaner" or "your line-height is too tight for this measure." Boring rules — but boring rules are exactly what taste is made of.

5. Color and contrast restraint. My instinct as a non-designer was always to add more — more colors, more accents, more "pop." The critique that stuck: restraint reads as confidence. One accent color, used sparingly, beats five fighting for attention. Learning to subtract was harder than learning to add, and it's the thing that made my work finally look calm instead of busy.

If you want the mechanical part — how to actually install and wire these up in Claude — I walked through that separately in how to add skills to Claude. This post is about what happened after I had them running.

What changed in my actual work

The point of all this, for me, was never to become a designer. It was to stop shipping App Store screenshots and app screens that quietly lost me users because they looked untrustworthy.

Here's the concrete change. I used to generate creative, look at it, feel a vague "eh," and ship it anyway because I couldn't articulate the problem. Now I have language. When I run a listing through Reverze's AI Style Remix and editor and it hands me a few directions, I can actually judge them — this one has a cleaner hierarchy, that one's leaning on a generic gradient, this third one has the confident restraint I want. I'm no longer picking at random and hoping. I'm directing.

That's the real unlock. AI can rebuild your screenshots and generate new directions all day, but the person choosing between them still needs an eye. Training mine is what turned "the tool made options" into "I shipped the right one." The generation got faster; my judgment got better; and those two things together are what actually moved my conversion numbers.

The honest limit (taste still takes reps)

I promised honesty, so here's the ceiling. None of this made me a designer overnight, and anyone selling that is lying to you. Claude can tell me why something's off, but I still had to look at hundreds of examples for the lessons to become instinct. The skills are a fast, patient teacher — they are not a shortcut around the reps.

There's also a trap: it's easy to lean on the critique so hard that you never develop your own opinion, just outsource it every time. The point is the opposite. I use the feedback until I can predict it, and then I need it less. If you're still asking Claude "is this good?" a year from now with no instinct of your own, the skill failed you — or you leaned on it instead of learning from it.

But for someone who genuinely couldn't design — who felt the wrongness and couldn't name it — this was the closest thing to a mentor I've had. It didn't give me talent. It gave me vocabulary, and vocabulary is where taste starts.

Try it on your own work

The fastest way to feel this is to run your own app through it. Paste your App Store URL into Reverze, let the Style Remix hand you a few directions, and then — instead of picking at random — try to say out loud why one is better than the others. That sentence you struggle to form is exactly the design sense worth building. Do it enough times and it stops being a struggle.

---

Reverze turns App Store creative production from a multi-day design workflow into minutes — paste a URL, rebuild your screenshots, generate new directions, and export production-ready assets. Explore the free tools or start in the app.