Case Study: Open Network Digital Company streamlines e-commerce product categorization with Netscribes and GPT

Open Network Digital Company - Customer Case Study

The open network digital company, India's first open network digital commerce platform, was challenged by inaccurate product categorization caused by seller-provided descriptions. Their initial manual process and subsequent NLP-based machine learning model required constant human quality control and frequent, costly retraining to manage an evolving product taxonomy. Vendor Netscribes was engaged to address this issue.

Netscribes implemented a solution using the GPT-3.5 language model with Few-Shot prompt loading techniques. This approach utilized dynamic taxonomy and product information to improve categorization. The new system streamlined model upgrades and drastically reduced retraining time from two weeks to just two days, resulting in substantial cost savings and enhanced operational efficiency for the open network digital company, despite a minor dip in accuracy.