Methodology

How We Build These Guides

Every color pairing, fit recommendation, and dress-code note on Clad9 traces back to a real, hand-curated relationship — not something an AI model guessed on the spot. Here's exactly what's real, what the AI does, and how we check its work.

1. The data is curated, not generated

Every color on Clad9 carries a real hex code, family, undertone, and seasonal-palette classification. Every color-pairing relationship, every garment-to-body-shape fit recommendation, and every occasion's dress code is a specific, hand-curated entry in our database — grounded in standard color theory (complementary, analogous, and neutral-anchor relationships) and established garment-fit and dress-code conventions used across the fashion and styling industry.

None of that structure is invented by a language model. If two colors are shown paired on Clad9, that pairing was curated as a real relationship before any AI wrote a single word about it.

2. AI writes the explanation, on top of that data

Once a pairing or fit relationship is curated, we use an AI model (Claude, via AWS Bedrock) to write the plain-English explanation of why it works — the part you read on the page. The AI is never asked to invent the relationship itself, only to explain one that's already been established.

3. Every draft is checked before it publishes

Generated explanations are compared against every other explanation already on the site using an embedding-similarity check. If a new draft reads too close to something we've already published, it's held back and rewritten rather than shipped as filler — so what you're reading was written specifically for that pairing, not lightly reworded boilerplate.

4. Not a substitute for a real stylist

This is general styling guidance grounded in real conventions, meant as a fast, free starting point — not personalized advice from a human stylist who's seen your wardrobe, your coloring, or your body in person. Use it as a foundation, not the final word.