<![CDATA[This study examined associations among Human-Like Interaction, Perceived Empathy, and Customer Loyalty in the use of conversational artificial intelligence (AI) for Indonesian e-commerce. A cross-sectional survey collected questionnaire responses from 300 purposively selected consumers who had experience with e-commerce chatbots or virtual assistants. The analysis used partial least squares structural equation modeling (PLS-SEM) in SmartPLS 4 to assess 18 indicators across three reflective constructs. The reported measurement properties met the specified reliability and validity thresholds. The reported PLS-SEM results indicated positive direct associations among Human-Like Interaction, Perceived Empathy, and Customer Loyalty, alongside a positive specific indirect association through Perceived Empathy. However, discrepancies between the structural model and the direct-effects table prevent a definitive interpretation of the estimated effect magnitudes and mediation type. However, coefficients in the structural diagram differed from those in the direct-effects table, so the magnitudes and type of mediation could not be confirmed from the reported outputs. The study distinguishes perceived conversational naturalness from perceived understanding and company-directed loyalty intentions. These findings concern self-reported perceptions and intentions rather than observed repurchase behavior. Its cross-sectional design and purposive sample preclude causal conclusions and population-wide generalization. The results provide a basis for evaluating interaction quality and perceived understanding in chatbot-supported customer service.]]>
Copyrights © 2026
