Drivers and Necessary Conditions for Chatbot Acceptance in the Insurance Industry. Analysis of policyholders' and professionals' Perspectives
Jorge de Andrés-Sánchez Jaume Gené-Albesa a Social and Business Research Laboratory,University Rovira i Virgili,Reus,Spainb Business Management Department,University Rovira i Virgili,Reus,SpainJorge de Andrés-Sánchez holds a Ph.D. from the University Rovira i Virgili (Spain),where he serves as a professor and researcher. He specializes in applying soft computing to actuarial,economic,and social challenges. His primary research focuses on technology acceptance and the ethical implications of emerging technologies in society. He has published extensively in high-impact international journals addressing social and business topics.Jaume Gené-Albesa earned his Ph.D. from the University Rovira i Virgili (Spain) and is a professor of Marketing in the Department of Business Administration at the same institution. He is an expert in digital commerce and marketing and has worked as an independent consultant for numerous firms. He has contributed to various prestigious international journals on marketing and commerce.
DOI: https://doi.org/10.1080/10919392.2024.2435118
IF: 2.2368
2024-12-04
Journal of Organizational Computing and Electronic Commerce
Abstract:The adoption of conversational robots (chatbots) for customer service is expanding across many industries. In the insurance sector, where customer interactions are essential to the use of policies, understanding chatbot acceptance is particularly relevant. This study explores the factors and conditions influencing the acceptance of chatbots for insurance policy management via the unified theory of acceptance and use of technology (UTAUT) framework. The analysis is conducted on two groups: ordinary policyholders and policyholders who are also industry professionals. The explanatory factors evaluated are performance expectancy, effort expectancy, social influence, and trust. The findings indicate that effort expectancy, social influence, and trust positively impact the behavioral intention to use chatbots. Additionally, all the variables are found to be necessary for acceptance. The structural equation model assessment reveals that professional status do not moderate the relationships between explanatory variables and behavioral intention; however, professionals demonstrate a greater intention to use chatbots. Among ordinary policyholders, effort expectancy has the largest effect size on acceptance. For professionals, trust and performance expectancy are the most impactful explanatory variables, with very large effect sizes. These results emphasize that while all variables are essential for acceptance, the relative importance of each variable varies between policyholders and professionals, offering insights for implementing chatbot solutions effectively within the insurance sector.
computer science, information systems, interdisciplinary applications