UX library and token audit

An audit of the UX library, components and design tokens, and the case for cleaning them up.

Tokens had to be clean before AI-assisted workflows could be production-ready.

My role

Auditor and author of the findings and stakeholder video.

Measured Storybook coverage, colour token sprawl, and spacing and radius values.
Split the remediation backlog into mechanical fixes and judgement calls.
Presented the findings with the numbers up front, to show stakeholders the scale.

A remediation backlog, split into mechanical fixes and judgement calls, and the case for fixing tokens before scaling AI workflows.

Researched: A scan of both code repositories: every colour, token, component and CSS class, plus Storybook coverage and the spacing and radius values in use. It showed how far the code had drifted from the design tokens.
Iterated: The raw inventory became a worklist. Every colour was mapped to a proposed scale and sorted into adopt, replace, review or decide. Dimensions were added from the same scan and clearly marked as partial until the scanner is widened.
Delivered: Two working tools that read the same scan, so they cannot disagree, and a stakeholder video that led with the findings to show the scale of the problem.
In short: Clean tokens are what make AI-assisted design and build production-ready. This made the clean-up measurable and split it into mechanical fixes and real decisions.

Looking for a product and UX/UI designer for a contract role? I'd like to hear about it.