just stumbled onto this approach using claude to run a proper design critique on generated components. instead of just accepting the first pass, it treats the output as a draft that needs active refinement thru iterative feedback loops. i usually just dump my
Figma screenshots into the chat and ask for a breakdown based on
usability heuristics and accessibility standards. it makes the process feel less like magic and more like a structured peer review session.
>the goal is to move beyond simple prompt engineering and focus on deep analysis.it helps catch those weird spacing errors or broken alignment issues that often slip through when u're relying on generative tools alone.
it basically turns the ai into a senior designer sitting next to you. i wonder if this will eventually make manual prototyping feel
obsolete much faster for everyone. does anyone else use specific system prompts to force this kinda critical thinking, or do u just wing it w/ natural language?
link:
https://uxplanet.org/claude-design-critique-a-better-way-to-improve-ai-generated-ui-7efa1064bf8a?source=rss----819cc2aaeee0---4