ZKH Group
Industrial product search
Buying required exact SKU codes. Now, AI agents identify products from natural language descriptions across 16 scenarios.
- AI optimization across 120 procurement sub-scenarios
Staff manually hunted for vendors to match precise specs. Gen AI now discovers suppliers and unlocks PDF catalog data instantly.
A B2B e-commerce platform transforming industrial supply chains, offering over 700,000 products to more than 1,000 manufacturing businesses.
Industrial buyers require precise product specifications, but staff had to manually search for items and often failed to locate the right vendors....
“Each of our technology projects drives one or more of these three impacts: enhance the customer experience, improve productivity, and boost governance. Our AI strategy executed through Google Cloud supports all three.”
B2B e-commerce marketplace for industrial, MRO, and business supplies.
Cloud computing services, AI infrastructure, and data analytics platforms for enterprises.
Related implementations across industries and use cases
Buying required exact SKU codes. Now, AI agents identify products from natural language descriptions across 16 scenarios.
Sourcing teams sifted through fragmented data to find parts. An AI assistant now scans datasheets to identify alternates instantly.
Reps relied on memory to pitch 30+ SKUs per brand. AI now analyzes local trends to generate dynamic "must-stock" lists.
Buying required exact SKU codes. Now, AI agents identify products from natural language descriptions across 16 scenarios.
Text matching failed on complex queries in a 50k catalog. Now, LLMs interpret intent to rank products by likelihood to convert.
IT staff logged in every six hours to check 30 systems. GenAI now automates monitoring, saving 3,000 hours and ending downtime risk.
Scouring foreign technical sources took months. AI now translates and analyzes materials in minutes.
Sonar review meant checking raw footage frame by frame. AI now scans hours of data to pinpoint targets for recovery.
Fragmented server data bottlenecked scaling. A unified model cut hardware SKUs by 50% and eliminated dozens of manual processes.