A Human-Centered Decision Support System in Customer Support

In customer support, employees have to handle a high number of customer e-mail inquiries every day. With an increasing variety of products, the number of different possible customer problems increases significantly. The employees face a sheer unmanagable task to distinguish between the different customer problems. In order to improve customer satisfaction and reduce employee stress and workload, we present a human-centered and artificial intelligence-based decision support system for employees in customer support. The system can be deployed towards classifying customer e-mail inquiries into pre-defined categories. For this purpose, we utilize a natural language processing pipeline around a recurrent neural network returning to the employees the best-fitting categories for incoming e-mails.