The Challenge
Rebel wanted to develop chat bots that would not only assist their internal customer service representatives but also enhance their customer’s experience while lowering costs. However, before Rebel’s chatbot could assist customers, they needed to first enrich its knowledge-base and train the chatbot by utilizing the expertise of their customer service agents.
How We Solved It
Sphere implemented a full-cycle chatbot development approach for Rebel through language parsing, syntax parsing, and machine learning technologies. The original XML interface of the chatbot was enhanced by using React.js and Angular.js for the web-facing application, and Spring Boot and JavaScript for the back-end. The new interface successfully removed the “rough edges” of the chat bot, and allowed it to assist both internal representatives and Rebel’s customers.
Sphere and Rebel also grew the knowledge database to provide quicker and more tailored answers. For example, if a customer asked about the status of their loan, Rebel needed their chatbot to first ask for the loan ID and then open the back-end application to contextualize the loan ID.
Machine learning and an open NLP framework were used to improve the chatbot’s ability to learn and interact with customers. Once the chatbot was completed, Sphere provided full maintenance and technical support.
The Results
Through our approach, Rebel was able to achieve their business goals by reducing costs and creating a competitive advantage by adopting the technology that will play a key role in the future of customer support.
