How I Use Claude as a Design Partner

Most designers I talk to are either not using AI at all, or they're using it like a search engine.

They ask vague questions. They get generic answers. They walk away unimpressed and assume Claude just isn't that useful for design work.

I used to feel the same way. Then I changed how I set it up.

The setup is the work

The outputs you get from Claude are almost entirely determined by what you give it before you ask for anything.

Vague input produces vague output. Specific input produces specific output. This sounds obvious but most people skip the setup entirely and go straight to the prompt — then wonder why the results feel generic.

I started treating Claude the way I'd treat a junior designer: someone capable, fast, and willing — but who needs direction. Not a genie. A collaborator who needs context.

That shift changed everything.

Give it structure, not adjectives

The first thing I do at the start of any design session is drop in a structured snapshot of the design system I'm working with.

{
  "colors": {
    "primary": "#2216FF",
    "canvas": "#FFFEFC"
  },
  "spacing": "8px grid",
  "radius": 12,
  "font": "DM Sans",
  "minBodySize": 16
}

{
  "colors": {
    "primary": "#2216FF",
    "canvas": "#FFFEFC"
  },
  "spacing": "8px grid",
  "radius": 12,
  "font": "DM Sans",
  "minBodySize": 16
}

{
  "colors": {
    "primary": "#2216FF",
    "canvas": "#FFFEFC"
  },
  "spacing": "8px grid",
  "radius": 12,
  "font": "DM Sans",
  "minBodySize": 16
}

Not a PDF. Not a mood board. Specific values.

"Modern and clean" is meaningless to a model. radius: 12 and gap: 24px are not. The more concrete the input, the less correction you'll need on the other end.

Codify your taste

Claude doesn't carry memory between sessions. Which means without some kind of preferences file, you'll spend the first few minutes of every conversation correcting the same mistakes.

I keep a short markdown file with my recurring design rules. Things like: always use an 8px grid, buttons at 12px radius, no drop shadows on cards, icons at 1.5px stroke weight. I paste it in whenever I'm doing design work.

It sounds tedious. It saves a lot of time.

Think of it as writing down your taste so you don't have to re-explain it every time.

Learn enough to speak the language

You don't need to be a developer. But a surface-level understanding of HTML and CSS will change how useful Claude is to you.

Claude produces code. If you can read it — even loosely — you can spot when something's off and correct it precisely. Without that, you're evaluating by eye and guessing at fixes.

The same goes for feedback. "Make it pop" doesn't translate. "Increase font-weight to 600" does. CSS is a shared language between you and the model. The more fluently you speak it, the less gets lost in translation.

Work in layers

The other big mistake I see is writing one giant prompt that tries to specify everything at once.

You lose control of the output. And when something's wrong, you don't know which part caused it.

What works better is working in layers. Start broad — "build me a card layout." Then refine a section. Then refine an element within that section. Then adjust a single property.

Each round narrows the gap. Small, directed changes keep you in the driver's seat. By the time you get to the detail level, what you're asking for is so specific that Claude almost can't get it wrong.

Save your versions

Claude has no memory. Which means if you iterate past something you liked, it's gone — unless you saved it.

Before any significant change, I label and copy the current state. Even pasting it into a Notion block titled v2-spacing is enough. Then when I want to go back, or combine elements from two different directions, I have the pieces.

"Go back to v2 but use the spacing from v3" is a powerful instruction. But only if you've been saving as you go.

Without versions, your best iteration is one wrong prompt away from being lost.

Art direction, not magic prompts

There's a version of using Claude that feels like trying to wish for the right output. You craft the perfect prompt, hope for the best, and adjust from there.

That's not how I think about it anymore.

The better mental model is art direction. You know what you want. Claude is fast and capable but has no taste of its own. Your job is to set the direction clearly, correct course precisely, and stay in control of the vision.

The best results come from good setup, not magic prompts.

Once I started thinking about it that way, the outputs stopped feeling generic. They started feeling like mine.