No open-access location is known for this record yet.
No open-access location is known for this record yet.
Despite the rapid integration of AI-powered services into e-retailing, many firms continue to face challenges in meeting customer expectations and sustaining long-term usage. Addressing this issue, the present study investigates the determinants of customers' continuance usage intention by developing and empirically testing an extended Expectation-Confirmation Model (ECM). A snowball non-probability sampling method was employed to collect data from 542 Vietnamese e-retail customers, and Partial Least Squares Structural Equation Modeling (PLS-SEM) was used to validate the research model and examine the proposed relationships. The findings reveal a sequential mechanism in which AI-powered service quality significantly enhances customer experience (O = 0.770, p < 0.001), perceived usefulness (O = 0.161, p < 0.001), and perceived problem-solving ability (O = 0.695, p < 0.001). These cognitive and experiential evaluations subsequently strengthen expectation confirmation and customer satisfaction. Notably, customer satisfaction and perceived usefulness emerge as the strongest predictors of continuance usage intention (O = 0.371 and O = 0.279, respectively; p < 0.001), with customer experience playing a central role in this process. This research extends the ECM by integrating AI service performance, multi-dimensional customer experience (encompassing hedonic aspects such as enjoyment and engagement, as well as recognition aspects such as feeling valued and understood), and perceived problem-solving ability into a unified framework tailored for the e-retailing context. The findings highlight an important managerial implication: sustaining users' continued engagement with AI-powered services requires simultaneous investment in both technical performance and the quality of the customer experience and satisfaction these systems generate.
Loading…
Stored only when both papers are in the corpus — 5 citations exist upstream.
Papers that cite the same references — the relational fallback while embeddings are still backfilling.
Imported from OpenAlex (CC0) — nothing here is generated.