AQuA: Automatic Quality Analysis of Conversational Scripts in Real-time

Call center is a point of contact between customers and the organization. A service agent represents the organization in one way and act as the face of the organization. The quality of service provided to a customer by a service agent strongly determines the reputation of an organization. Hence, good quality service is of utmost importance for an organization’s continual growth in the present dynamic market situation. In order to streamline the process of providing exceptional service, organizations have placed human experts to evaluate the quality of calls. However, evaluation through human experts is labour-intensive, limited to few random calls and expensive. In this paper, we propose a Quality AI Assistant which actively monitors all agents’ interactions and provides feedback to improve call quality in real-time. We also propose a model to predict the customer satisfaction. We apply Deep Learning based approach to evaluate the quality of calls on different organizational compliance aspects and predict the customer satisfaction. We evaluated the performance of our system on an in-house contact centre dataset of finance domain. The experiment results demonstrate the effectiveness of the approach which is quite promising.