Next Generation of Cleansed Interactive Contact Centre for Efficient Customer Recognition

Contact centres (CCs) are one of the main touch points by which customers can contact an organization. They are also one of the sources of input to Customer Relationship Management (CRM) systems that enable an organization to better manage their relationship with their customers. In today’s competitive world, it is important that organizations have efficient CRM systems to ensure that the loyalties of their customers remain with them. As an indication of their importance, in the literature, various predictive and innovative CRM approaches have been proposed. However, for the better management of their operations, apart from only focusing on the satisfaction and loyalty of customers, organizations should also focus on enhancing and maintaining the satisfaction of their Customer Service Representatives (CSRs), who are the front line staff working in CCs. For this, they need to develop innovative tools that will assist the CSRs in doing their job efficiently, thereby increasing their work satisfaction. In spite of recognizing the importance of such systems in the literature, no formal approaches have been proposed that help to achieve these aims. This is in contrast to the numerous predictive and innovative CRM approaches proposed that focus on increasing the satisfaction of both existing and new customers. Thus, to address this gap, this thesis proposes innovative systems that assist CSRs to undertake their tasks efficiently. This will also lead to an improved and positive customer experience as a result of their interaction with the CSRs. In order to address the problem and realize these benefits, the Cleansed Interactive Customer Recognition Framework (CICRF) is proposed. CICRF has two broad aims, first is the deduplication of the CC's database (DB) using a hybrid technique of crowdsourcing and machine learning and the second is the development of an interactive framework for customer recognition in CCs using hybrid heuristic classical techniques. For each aim of CICRF, various computational phases with several methods and metrics are proposed. CICRF is also evaluated on synthetic datasets. The results show that CICRF assists in the cleansing of a CC’s DB as well as in the recognition of the correct customer