In many situations, a call centre agent acts as an interface between the customer and a knowledge. Research has shown that Customer Satisfaction correlate positively with the quality of the underlying Knowledge Bases. But creating and maintaining Knowledge Bases is expensive and time-consuming, therefore we are working on ways to optimize and automate the way knowledge and insight is created, consumed and shared across the support channels that serve end customers.
Xerox research is working on a tool for agents which analyses and automatically selects Knowledge Base entries based on customer enquiries.
The first step of this process is to apply Natural Language Processing techniques to disambiguate customers’ request and trigger the appropriate routing and find the most suited answer in the knowledge base.
The second step is to apply Natural Language Generation for presenting the answer retrieved from the knowledge base in natural language.