What challenges did ZKH Group Limited face before AI implementation?

Traditional MRO procurement depended on exact SKU lookups through search engines and customer service channels. ZKH needed a robust multi-region infrastructure with high resilience, fast international communication, and no single points of failure as it expanded across APAC and the United States.

What was the impact of the AI solution?

ZKH said the AI-powered solution reduced RFQ time. The company also said the approach streamlined procurement processes, accelerated global expansion, and improved operational efficiency and cost effectiveness.

Source

AlibabaProductionEvidence: Medium55/100

ZKH builds industrial procurement AI agents with Qwen and Model Studio on Alibaba Cloud

Use case typeAI model trainingUpdated 5 days ago

ZKH Group Limited is an MRO procurement platform that modernized legacy procurement workflows and expanded globally with Alibaba Cloud. The company built its own industry model in 2024 using the open-source Qwen model on Alibaba Cloud and fine-tuned it with e-commerce data to create a specialized expert model for daily MRO scenarios. ZKH deployed multiple AI agents to standardize materials, provide interactive services, compare supplier prices, and recommend products from natural-language requirements. The broader platform architecture also included ECS GPU Instances, ACK, ASM, OSS, PolarDB, MaxCompute, and Cloud Enterprise Network to support AI, storage, data processing, and international connectivity.

Organization: ZKH Group Limited
Industry: Logistics
Location: China
Published: July 2026

Reported outcomes

Strategic outcomes

Innovation & cultureReduced RFQ timeInnovation & cultureExpanded operations beyond APACCost efficiencyImproved operational efficiency and cost effectiveness

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: ZKH Group Limited
Provider: Alibaba
Maturity: Production
Linked source: Alibaba Cloud Customers

ZKH deployed multiple AI agents to standardize materials, provide interactive services, compare supplier prices, and recommend products from natural-language requirements

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 2 of 2

1AI model training
2Supply chain optimization

Traditional MRO procurement depended on exact SKU lookups through search engines and customer service channels.
ZKH needed a robust multi-region infrastructure with high resilience, fast international communication, and no single points of failure as it expanded across APAC and the United States.

ZKH used Alibaba Cloud Model Studio to access Qwen-Max and also built and fine-tuned its own expert model with Qwen on ECS GPU Instances.
It deployed AI agents for material standardization, interactive MRO support, supplier price comparison, and product recommendation.
It ran the application on ECS, ACK, ASM, OSS, PolarDB, MaxCompute, and Cloud Enterprise Network.

Technologies

Alibaba Cloud Model Studio Qwen Qwen-Max Elastic Compute Service Container Service for Kubernetes Service Mesh Object Storage Service PolarDB MaxCompute Cloud Enterprise Network

ZKH said the AI-powered solution reduced RFQ time.
The company also said the approach streamlined procurement processes, accelerated global expansion, and improved operational efficiency and cost effectiveness.

Architecture

ZKH built a specialized MRO expert model by fine-tuning Qwen with e-commerce data on ECS GPU Instances, then deployed multiple AI agents and supporting cloud infrastructure across ACK, ASM, OSS, PolarDB, MaxCompute, and Cloud Enterprise Network.

Sources & evidence1

Evidence: Medium55/100Evidence strength