AI-Powered Product Categorization & Data Extraction:

Automating Taxonomy for 80,000+ SKUs

$500M CONSUMER PRODUCTS COMPANY

Background

  • Two $250M Consumer Products & Apparel businesses were merging & consolidating ERP systems
  • Client had over 80,000 SKUs that needed to be categorized into a new, consolidated six-level product hierarchy
  • Valuable master data was embedded in part descriptions but difficult to extract manually
  • Manual categorization and data extraction processes were time-consuming and error-prone
  • Client sought an automated solution to streamline product taxonomy and master data management

Contributions

  • Developed a custom-trained GPT (Generative Pre-trained Transformer) model specifically tailored to the client’s product data and hierarchy
  • Leveraged advanced natural language processing techniques to enable accurate SKU categorization and master data extraction
  • Utilized the AI model to optimize the product categorization process, resulting in a more efficient consolidation of ERP systems post-merger.
  • Integrated the model into a user-friendly interface for easy access and utilization by the client’s team

Results

  • Categorized 80,000+ SKUs into the six-level product hierarchy with high accuracy
  • Extracted valuable master data from part descriptions, enabling enrichment of product records
  • Significantly reduced the time and manual effort required for product categorization and data extraction
  • Improved data quality and consistency across the product catalog
  • Laid the foundation for enhanced product analytics