AI is becoming part of every designer's toolkit, but I believe its biggest value isn't generating interfaces. It's helping solve complex, repetitive problems faster so we can focus on the work that actually requires critical thinking.

I recently put that into practice while tackling a design system issue that would have taken weeks, if not months, to audit manually.

Challenge

Our design tokens in Figma were no longer aligned with the tokens in Storybook. Some HEX values were inconsistent, semantic tokens pointed to different primitive colors, and a few semantic tokens weren't even connected to the correct primitive. Auditing every token by hand wasn't realistic, so I looked for a better solution.

Steps I took with AI

Using Cursor, I built an AI-powered auditing tool that compared our Figma tokens against Storybook and automatically flagged inconsistencies. I started by visually displaying the primitive colors from Figma inside an AI-built interface so I could scan the palette quickly and spot issues at a glance. I then did the same with the primitives from Storybook, but this time I added indicators to show exactly which HEX values did not match what was defined in Figma. That made it much easier to identify where the two sources had drifted apart without having to compare everything manually.

Once I cleaned up the HEX values for the primitives, I moved on to creating a semantic comparison between Figma and Storybook. This part was more challenging because the semantics did not match at all. Even so, I was able to surface every semantic token that was either not linked to a primitive or was using a made-up HEX value. For those cases, the tool proposed the closest primitive color it thought should be used instead. I manually verified every proposal to make sure it still felt on brand and aligned with the design system.

So instead of spending my time hunting for inconsistencies, I could focus on validating the results and making informed design decisions.

Result

The tool transformed what would have been a tedious manual audit into a fast, reliable workflow. It saved me weeks of work, improved confidence in the design system, and highlighted issues that would have been easy to overlook.

This project reinforced something I've come to appreciate about AI. Its greatest strength isn't replacing designers. It's eliminating repetitive work so we can spend more time solving meaningful problems. Sometimes the best design solution isn't another Figma component. It's building the right tool for the job.