Determinants of customer satisfaction in chatbot use : evidence from a banking application in Turkey

Purpose: This study aims to investigates customer satisfaction from the use of bank chatbots and the effect of perceived trust in chatbots and banks' reputation on customer satisfaction. Design/methodology/approach: A survey is conducted in Turkey involving 240 customers who experienced banking transactions using a chatbot. Partial least squares structural equation modeling (PLS-SEM) is used to investigate the relationships between the variables. The data were analyzed using SPSS 21 and SmartPLS programs. Findings: Perceived performance, perceived trust and corporate reputation significantly affect customer satisfaction with chatbot use. Customer expectations and confirmation of customer expectations have no direct impact on customer satisfaction, but customer expectations positively affect perceived performance. Customer expectations exert an indirect influence on customer satisfaction through perceived performance. Perceived performance has a positive impact on the confirmation of customer expectations, but customer expectations do not significantly impact the confirmation of customer expectations. Research limitations/implications: This study relies on a limited number of participants. Moreover, its sample is not representative of the target population due to the convenience sampling technique. Even if the results may not be generalized to the entire population of Turkey, they reflect the reality of emerging markets with relatively high technology sensitivity and a young population. Practical implications: The results provide new insights regarding banking service delivery channels, which may be of interest to professionals, academics, banks' top management, product development teams, design teams and customer satisfaction units. Social implications: This study is believed to help the community make their lives easier by providing them with knowledge and awareness about chatbots. Originality/value: This study extends expectations confirmation theory's predictions to chatbot use in banking.