AI Taxonomy Engine
Lesuto's AI taxonomy engine reads product titles and descriptions, then extracts color, material, style, room, eco-certifications, and 12+ more attributes automatically — with disambiguation that knows "oak finish" is a color and "solid oak" is a material.
Now Available
Rule-based disambiguation with AI fallback — 600+ rules, 17 universal facet types, and contextual analysis that prevents misidentification.
Analyzes product titles and descriptions to extract color, material, style, room, audience, condition, season, eco-certifications, and 10+ more attributes — with no manual tagging.
Distinguishes ambiguous terms using 600+ contextual rules. "Oak finish" is correctly tagged as a color (brown), while "solid oak" is tagged as wood material. Confidence scoring flags edge cases for review.
Every extraction includes a confidence score and evidence trail. Only high-confidence results are auto-applied. Medium and low-confidence extractions are flagged for human review.
Each extracted attribute shows exactly what text triggered it, which rule matched, and the confidence reasoning — full transparency for every classification decision.
"Faux leather" and "vegan leather" are correctly identified as synthetic — distinguished from genuine materials using negator detection. Critical for accurate product listings and legal compliance.
A single product can yield 8+ attributes simultaneously: color, material, style, room, category, audience, eco-certification, season, and price range — all from title and description text.
How It Works
Products enter from any source — supplier uploads, imports, pricing sheets, or the product wizard. The engine runs automatically.
Add products via supplier channels, imports, pricing sheets, or the product wizard. The taxonomy engine activates automatically on every new or updated product.
The engine runs 600+ disambiguation rules and keyword maps first. Ambiguous or low-confidence cases escalate to Claude AI for contextual analysis.
High-confidence extractions are applied as product facets immediately. 17 universal facet types — color, material, style, room, audience, eco-friendly, and more.
Products become filterable, searchable, and categorized across storefronts, the SDK, WordPress, and the Hub — improving product discovery and conversion.
Your Single Source of Truth
Browse tree classification, bullet point generation, backend search terms, A+ content-ready descriptions, and ASIN mapping.
Wayfair-specific category taxonomy, dimension formatting, material attributes, and assembly instruction flags.
Tag generation, materials list, craft-friendly descriptions, holiday and seasonal tagging, and shop section mapping.
Walmart taxonomy classification, rich content attributes, shelf description, shipping weight mapping, and category-specific requirements.
Why This Matters
Coming Soon
The same extraction engine that powers internal tagging will expand to auto-format products for Amazon, Wayfair, Etsy, and Walmart.
Your Single Source of Truth
Browse tree classification, bullet point generation, backend search terms, A+ content-ready descriptions, and ASIN mapping.
Wayfair-specific category taxonomy, dimension formatting, material attributes, and assembly instruction flags.
Tag generation, materials list, craft-friendly descriptions, holiday and seasonal tagging, and shop section mapping.
Walmart taxonomy classification, rich content attributes, shelf description, shipping weight mapping, and category-specific requirements.
Every product you add to Lesuto gets AI-powered attribute extraction. Color, material, style, room, eco-certifications — all extracted and applied automatically. No extra cost.
Get Started Free