How AI Helped Us Make Sense of Hundreds of Components

Back to Devoured - July 21, 2026

DEVOURED

How AI Helped Us Make Sense of Hundreds of Components

Design frontendai Clearleft

Clearleft used conversational AI to audit an unmanaged library of hundreds of legacy web components for a UK university.

What: Design agency Clearleft managed a chaotic component inventory by using LLM-assisted workflows to categorize and document years of accumulated code. Instead of drafting a rigid specification, the team iterated conversationally to classify components without manual line-by-line inspection.

Why it matters: This demonstrates a practical alternative to massive documentation projects; using LLMs to make sense of 'code rot' allows teams to regain control over legacy systems incrementally.

Deep dive

Use conversational prompting to label component categories by summarizing file structures rather than relying on existing metadata.
Feed component code snippets to an LLM to identify duplicates or common patterns in UI styles.
Leverage AI to generate standardized documentation or JSDoc comments for undocumented components.
Validate machine-generated groupings with human oversight to ensure consistency across the design system.

Decoder

Design system: A collection of reusable components, standards, and patterns that guide the creation of consistent user interfaces across a product or organization.

Original article

A UK higher education institution had years of organically grown web components with no clear inventory, so a team built an AI-assisted audit tool through rapid, conversational iteration rather than a fixed spec.

Explore all 43 articles in today's digest