User interactions with chatbot interfaces vs. Menu-based interfaces: An empirical study

Abstract

Rapid advances in Natural Language Processing (NLP) are transforming customer service by making it possible to create chatbot applications that can understand users’ intents and response in a human-like manner. Chatbots promise to enhance customer experiences by creating more personal customer interactions than those afforded by traditional menu-based web applications. But are chatbots always superior to more traditional user interfaces (UI)? This study seeks to understand the differences in user satisfaction with a chatbot system vis-a-vis a menu-based interface system, and identify factors that influence user satisfaction. Grounded in the self-determination theory, the research model proposed here focuses on the effect of chatbot use on perceived autonomy, perceived competence, cognitive load, performance satisfaction, and system satisfaction. An experimental study was conducted, and data were analyzed using Partial Least Square Structural Equation Modeling. The findings indicate that chatbot systems lead to a lower level of perceived autonomy and higher cognitive load, compared with menu-based interface systems, resulting in a lower degree of user satisfaction. Implications of these findings for research and practice are discussed.