<![CDATA[Artificial intelligence-powered chatbots have revolutionized e-commerce by providing personalized customer interactions, real-time support, and streamlined purchase processes. Despite their widespread adoption, sustained user engagement remains challenging, requiring deeper insights into cognitive, affective, and social determinants of long-term usage. This study addresses this gap by integrating the Technology Acceptance Model (TAM), Theory of Planned Behavior (TPB), and Expectation-Confirmation Theory Model (ECTM) into a comprehensive triangulated framework to examine user acceptance and continued purchase intention toward AI chatbots in online shopping. The research investigates direct effects of confirmation, information quality, perceived usefulness (PU), perceived ease of use (PEOU), attitude, and subjective norm on satisfaction, alongside satisfaction's mediating role in predicting continued purchase intention. Data were collected from 504 respondents with prior AI chatbot experience in online shopping via purposive sampling, using validated 6-point Likert scales. Partial least squares structural equation modeling (PLS-SEM) was conducted using SmartPLS 4. Results confirm that confirmation (β=0.178, p=0.037), information quality (β=0.269, p0.001), PU (β=0.152, p=0.005), PEOU (β=0.235, p0.001), and attitude (β=0.184, p=0.001) significantly predict satisfaction, which strongly influences continued purchase intention (β=0.868, p0.001). Subjective norm exhibited no significant effect (β=-0.003, p=0.954). Satisfaction fully mediates ECTM and TAM pathways, underscoring experiential confirmation and system quality's dominance over social influences in post-adoption behavior. Theoretically, this study validates an integrated model advancing post-adoption theory in AI contexts. Practically, findings guide e-commerce platforms to enhance chatbot retention by prioritizing information accuracy, usability, and expectation alignment rather than social norms.]]>
Copyrights © 2026
