Ecommerce ops · Production
Zalando builds a fashion knowledge graph with semantic web technologies to enable implicit concept search and vegan-aware filtering
The problem
Zalando's product data lacked the background knowledge to express implicit fashion concepts — for example, that the absence of wool, silk, and leather equals 'vegan' — making it impossible to answer customer queries like 'vegan clothes' from product attributes alone.
Workflow diagram · grounded in source
1
Expressivity gap identified
trigger
“the newly-opened Helsinki office wondered about the expressivity of our current product data. What do attributes like material construction or sport quality really mean”
2
Knowledge graph constructed
ai_action
“building a fashion knowledge graph for Zalando using Turtle, JSON-LD and Python. Semantic web technologies use knowledge graphs, technically known as named directed graphs, to provide background data held by humans in a machine-readable …”
3
Product data interpreted via ontology
ai_action
“an application using the knowledge graph can interpret the product data of wool, silk, or leather as not suitable for vegans, or only offer products that are vegan by excluding those items with materials made from animals”
4
Semantic search and filtered results surfaced
output
“Understand the word "vegan" in a search without ever expressing it in our product data explicitly, Offer special values in filters, such as "vegan", Show a page with knowledge about vegan clothing that includes articles on vegan fashion,…”
5
Business rules and link suggestions applied
routing
“Intelligently suggest links for further browsing. Apply business rules. For example, when a customer is browsing a particular brand, we might not suggest competing brands”
Reported outcome
The fashion knowledge graph enables Zalando to understand 'vegan' in search without it appearing in product data, offer it as a filter value, and show knowledge pages combining articles, outfits, and vegan-appropriate products from the catalogue — and to apply business rules such as suppressing competing brand suggestions.
Reported stack
TurtleJSON-LDPython
Frequently asked questions
What did this team achieve with this AI workflow?
The fashion knowledge graph enables Zalando to understand 'vegan' in search without it appearing in product data, offer it as a filter value, and show knowledge pages combining articles, outfits, and vegan-appropriate…
What tools did this team use?
Turtle, JSON-LD, Python.
How is this ecommerce ops AI workflow structured?
Expressivity gap identified → Knowledge graph constructed → Product data interpreted via ontology → Semantic search and filtered results surfaced → Business rules and link suggestions applied.