THE EFFECT OF CHATBOTS RESPONSE LATENCY ON USERS’ TRUST

Abstract

Chatbots are widely used as conversational agents and being designed using anthropomorphic design guidelines. However, response latency (response latency is the time it takes for a chatbot/person to provide a response immediately after receiving a message) as an anthropomorphic design cue in a conversational user interface has not been the subject of many studies. Even though the system's response latency has an undeniable effect on users' satisfaction and performance, the connection between users' trust and chatbots' response time is not addressed. A critical reason that executives are reluctant to implement chatbots for their businesses is the user adoption hesitancy. Customers and users are unwilling to engage with a chatbot because they do not trust chatbot. Therefore, this study used empirical data collected from chatbot users to investigate the effect of chatbots response latency on users’ trust – cognitive and affective trust. The results of this study suggest that dynamically delaying chatbot response increases users’ cognitive trust but has no significant impact on users’ affective trust. General sentiment analysis on chatbot users’ responses to an open-ended question that describes their experiences interacting with chatbots suggests that dynamically delaying chatbot response produces higher positive sentiment and trust sentiment than near-instant chatbot response. Other findings are discussed and some ideas for future research are also presented in this paper.