Unionthink Leverages AI to Transform Inquiry Management

Unionthink, a Japanese ISV specializing in solutions for small and medium-sized enterprises, faced an increasing volume of product inquiries to its customer center. The existing manual categorization process for these inquiries was becoming a bottleneck, hindering efficient responses and potentially impacting customer satisfaction. To address this challenge, Unionthink sought to streamline its processes, reduce classification workload, improve customer satisfaction, and gain better insights into inquiry trends.

To overcome these obstacles, Unionthink partnered with IBM, leveraging IBM Watson's Natural Language Processing Library for Embed and watsonx.ai. Together, they developed a custom language model trained on Unionthink's manually categorized inquiries. This model enabled the customer center to more effectively address inquiries through AI-powered category classification. Through rigorous testing and refinement, Unionthink identified the most accurate model and optimized its performance through modifications and tuning of the training data.

The collaboration extended to a proof-of-concept (PoC) where Unionthink explored AI-driven tasks such as keyword extraction from inquiries. This PoC solidified the potential of AI technology to resolve Unionthink's challenges and improve business efficiency. The results of this collaboration have set Unionthink on a path to transitioning from manual labor to AI-driven solutions.

Unionthink's successful implementation of AI for category classification has delivered significant value:

• Increased prediction accuracy: Achieved over 90% prediction accuracy for classifications, improving the speed and accuracy of inquiry routing.

• Enhanced business efficiency: Streamlined inquiry processing, reduced workload on customer center staff, and freed up resources for other tasks.

• Improved customer satisfaction: Faster and more accurate responses to inquiries, leading to a better customer experience.

• Data-driven insights: Improved ability to analyze inquiry trends, enabling better-informed decisions about FAQs and product improvements.