Industry 4.0's impact on the insurance sector is substantial, driving the growth of insurtech. One key technology is chatbots, which offer policyholders convenient policy management. This study explores policyholders' attitudes towards chatbots, specifically their use in actions such as claim reporting. The researchers aim to determine the overall acceptance level of chatbots for insurer communication, and analyze the influence of Technology Acceptance Model (TAM) variables (perceived usefulness and perceived ease of use), as well as trust, on attitude and behavioral intention. The context is crucial because policyholders, unlike prospective customers, have an established relationship with the insurer, adding a dimension of trust to the technology adoption decision. Understanding this acceptance is vital for insurers to effectively leverage chatbots and improve customer experience while maintaining trust.

The study grounds its analysis in the Technology Acceptance Model (TAM), integrating it with the Unified Theory of Acceptance and Use of Technology (UTAUT) and emphasizing the critical role of trust in the context of AI-powered technologies. Existing literature on TAM, UTAUT, and the significance of trust in the adoption of financial technologies and chatbots are reviewed. Previous research using TAM and UTAUT to model chatbot acceptance is discussed, along with studies highlighting the impact of trust on perceived usefulness and ease of use. The authors posit that trust encompasses cognitive (perceived effectiveness) and relational (confidence in the insurer's chatbot implementation) dimensions. This theoretical framework underpins the hypotheses tested in the study, examining the relationships between perceived usefulness, perceived ease of use, trust, attitude, and behavioral intention regarding chatbot usage in insurance processes.

Data were collected using a structured questionnaire administered in Spanish to a sample of insurance policyholders. An initial pilot test with industry professionals refined the questionnaire before its administration to a larger sample. The questionnaire measured behavioral intention (BI), attitude toward chatbots (ATT), perceived usefulness (PU), perceived ease of use (PEOU), and trust (TRUST). The sample profile, including gender, age, education, income, and the number of insurance policies, is described. The sample size was considered adequate based on established rules of thumb for Partial Least Squares Structural Equation Modeling (PLS-SEM). The research questions were addressed using descriptive statistics (t-tests and Wilcoxon tests) to analyze average intention and attitude and PLS-SEM to evaluate the structural model and test the hypotheses. Scale reliability and validity were assessed using Cronbach's alpha, composite reliability, average variance extracted (AVE), Fornell-Larcker criterion, and heterotrait-monotrait (HTMT) ratios. The predictive capability of the model was evaluated using Stone-Geisser's Q² and cross-validated predictive ability (CV-AP).

The results showed low average intention and attitude towards using chatbots for policy management. PLS-SEM analysis revealed that the proposed model provided a good fit and predictive capability. All three factors—trust, perceived usefulness, and perceived ease of use—significantly influenced attitude and behavioral intention. However, trust exhibited the strongest impact, directly affecting attitude and intention, and also mediating the influence of perceived usefulness and perceived ease of use on attitude and intention. The R² values for ATT and BI were 0.717 and 0.604, respectively, indicating substantial explanatory power. The Q² values confirmed good predictive power for all constructs. Hypotheses regarding the positive effects of trust, perceived usefulness, and perceived ease of use on attitude and behavioral intention were largely supported.

The findings indicate that customer acceptance of chatbots for insurance policy management is low, confirming previous research highlighting customer reluctance in this area. The study's key contribution is the identification of trust as the most influential factor driving acceptance, surpassing perceived usefulness and ease of use. This underscores the importance of building and maintaining trust in the context of AI-powered solutions in the insurance industry. The significant mediating effect of trust on the relationship between perceived usefulness/ease of use and behavioral intention highlights that trust is not merely an independent factor but also amplifies the effects of other factors. The results are consistent with the literature on technology acceptance and the unique challenges of using chatbots in contexts requiring high levels of relational trust like insurance.

This study confirms the importance of TAM constructs and trust in explaining insurance customers' acceptance of chatbots. Trust emerges as the most critical factor, underscoring the need for insurers to foster trust in their chatbot systems. Future research should investigate the influence of emotional trust, explore different chatbot functionalities and user demographics, and conduct longitudinal studies to track acceptance over time. Further research could also explore chatbot applications across different stages of the insurance lifecycle and across varying cultural contexts.

The study's limitations include its cross-sectional nature, limiting the ability to make inferences about long-term trends. The sample, primarily from Spain and consisting of individuals with higher education levels and professional status, may not be fully generalizable to other populations. Future research should address these limitations by including longitudinal data and broader sampling across various demographic and cultural contexts.