Understanding and Forecasting Chatbot Adoption: An SEM-ANN Methodology

Abstract

This study introduces an innovative and integrated research model, synthesizing components from the established Technology Acceptance Model (TAM) with key elements like content richness and personal innovativeness, essential for Chatbot effectiveness. TAM serves as a foundation to identify factors influencing Chatbot adoption. A notable aspect of this research is its focus on Chatbot utiliza-tion for educational purposes, primarily aimed at augmenting the efficacy of inter-action between teachers and students. Our model highlights the direct correlation between TAM's perceived usefulness and perceived ease of use, and user satis-faction. Additionally, flow theory is integrated as an external factor, examining user involvement and control over Chatbot use. Data for this study was gathered from 824 education professionals, including teachers, administrators, and stu-dents. A novel hybrid analysis approach combining Structural Equation Model-ing (SEM) and an Artificial Neural Network (ANN) based on deep learning was employed. The study also utilized Importance-Performance Map Analysis (IPMA) to evaluate the significance and effectiveness of each factor. The findings from ANN and IPMA analyses pinpoint user satisfaction as the key determinant of Chatbot adoption intention. Structural equation modeling of the data reveals significant impacts of perceived usefulness and ease of use on the intention to use Chatbot. Furthermore, user satisfaction and flow experience emerge as pivotal in enhancing Chatbot adoption intention. Theoretically, our model offers compre-hensive insights into factors affecting Chatbot usage intention, particularly con-sidering internet service factors at the individual level. Practically, these results can guide decision-makers and practitioners in higher education to prioritize spe-cific factors and shape their strategies accordingly. Methodologically, this re-search demonstrates the deep ANN architecture's efficacy in elucidating complex, non-linear relationships within the theoretical model. The overarching conclusion indicates a high demand for Chatbots in education, serving as a prevalent com-munication medium between teachers and students, and significantly facilitating information exchange.