Case Study: Unionthink improves inquiry classification efficiency with IBM watsonx.ai and IBM Watson NLP

Improving category classification for manual inquiries

Unionthink, a Japanese independent software vendor serving small and medium-sized businesses, faced rising product inquiries at its customer center and had been manually categorizing them to improve FAQs and analyze trends. To keep up with demand, Unionthink looked to IBM and its AI tools, including IBM Watson Natural Language Processing Library for Embed and IBM watsonx.ai, to reduce manual work and improve response quality.

IBM worked with Unionthink to build a custom language model trained on manually categorized inquiries, then refined it through model comparison, technical verification, and data tuning. The solution achieved more than 90% prediction accuracy for classifications, helping Unionthink improve efficiency, better understand inquiry trends, and support higher customer satisfaction.

Unionthink