Automated question answering for customer helpdesk applications

This paper describes a closed domain question answering system which can be used as the first step in automating a customer helpdesk of a commercial organization. Even though there has been an increasing interest in data-driven methods over the past decade to achieve more natural human-machine interactions, such methods require a large amount of manually labeled representative data on how user converses with a machine. However, this is a strong requirement that is difficult to be satisfied in the early phase of system development. The knowledge-based approach that we present here is aimed at maximally making use of the user experience available with the customer services representatives (CSRs) in the organization and hence not relying on application data. The approach takes into account the syntactic, lexical, and morphological variations, as well as a way of synonym transduction that is allowed to vary over the systems knowledgebase. The query understanding method, which is based on a ranking algorithm and a pattern writing process, takes into account the intent, context, and content components of natural language meaning as well as the word order. A genetic algorithm-based method is presented for regularly updating the ranking parameters to adapt to changes in the nature of users' queries over time. We present an evaluation of our system deployed in a real-world enterprise helpdesk environment at Exetel Pty Ltd., Australia.