The Impact of AI-Powered Service on Customer Continuance Usage Intention in E-retailing: An Extended Expectation Confirmation Model

Despite the global interest in AI-powered services, research has largely concentrated on mature markets, leaving a significant contextual gap for rapidly growing digital economies like Vietnam. Vietnam's e-commerce market is experiencing substantial growth, driven by a young, tech-savvy population and high internet penetration. Furthermore, Vietnam is a regional leader in AI adoption within e-commerce, making it an ideal setting to study consumer perceptions and interactions with AI services. Existing research on AI-powered services in e-retailing often focuses on specific AI tools in isolation, adopting a "point solution" approach that overlooks how customers form a holistic judgment about the overall effectiveness of an integrated suite of AI services encountered throughout their shopping journey. Marketing theory emphasizes that customer perceptions are shaped holistically across the entire customer journey, yet how customers form cumulative judgments about AI performance from sequential and varied interactions remains an unanswered question. A significant gap also exists in the understanding of the customer experience itself. Much of the current literature is rooted in utilitarian-focused technology adoption models like TAM, which prioritize cognitive outcomes such as perceived usefulness and fail to capture the multi-dimensional nature of customer experience, including critical affective and social dimensions. The "Experience Economy" highlights the importance of memorable interactions, but research on AI in e-retailing has been slow to incorporate this, paying limited attention to how holistic AI performance affects hedonic dimensions like enjoyment and engagement. Additionally, recognition-related dimensions, such as feeling valued and understood through personalization, are crucial for an experiential evaluation but remain underexplored. Finally, the evaluation of AI's functional aspect is often too broad; perceived problem-solving ability offers a more precise, competence-based judgment that captures AI's effectiveness in critical moments of need, and its effect on satisfaction and continuance usage intention is underexplored.

AI-powered customer services represent a broad category of services that leverage AI technologies to automate, enhance, and optimize business processes and customer experiences. In e-retailing, these services include chatbots, recommendation engines, virtual assistants, and visual search technologies, which are becoming integral to the customer journey. Chatbots, for instance, offer 24/7 immediate and personalized support, efficiently addressing inquiries and guiding customers. Recommendation engines utilize machine learning to analyze customer data and deliver personalized suggestions that drive engagement and loyalty. This comprehensive integration of AI is revolutionizing e-retailing by providing unprecedented convenience, personalization, and efficiency. To understand the impact of these technologies, a modern approach is needed to evaluate their performance, moving beyond traditional frameworks designed for human-to-human interactions. The unique characteristics of AI services, such as automation and predictive capabilities, demand a technology-centric evaluation. Recent studies highlight that the impact of AI-powered customer service on loyalty is mediated by customer satisfaction and perceived efficiency. This study conceptualizes AI-powered customer service by examining four fundamental attributes that determine its performance and effectiveness from the user's perspective: reliability, security, helpfulness, and interface design. Reliability refers to the AI's ability to deliver promised services accurately and dependably. Security focuses on protecting customers' sensitive data. Helpfulness reflects the system's responsiveness and effectiveness in delivering support and resolving problems. Interface design captures the ease, clarity, and efficiency of interacting with the AI. Collectively, these attributes form the backbone of effective AI-powered customer service, critical for driving favorable customer perceptions and fostering long-term loyalty.

Perceived usefulness and perceived problem-solving ability are critical cognitive evaluations that influence customer satisfaction and continuance usage intention in the context of AI-powered services in e-retailing. The study's findings indicate that AI-powered service quality significantly enhances perceived usefulness, which is a key predictor of continuance usage intention. Perceived usefulness refers to the extent to which a user believes that a particular technology will enhance their job performance or facilitate their tasks. In the e-retailing context, this translates to how well the AI service helps customers find products, make informed decisions, or complete transactions efficiently. Furthermore, the research highlights perceived problem-solving ability as a distinct and important attribute of AI-powered customer service. This attribute captures the AI's competence in effectively addressing customer inquiries and resolving issues, offering a more precise, competence-based judgment than general usefulness. The study found that AI-powered service quality also significantly enhances perceived problem-solving ability. Both perceived usefulness and perceived problem-solving ability, alongside customer satisfaction, are crucial drivers of customers' intention to continue using AI-powered services. This emphasizes that while the AI must be functional and efficient, its ability to effectively solve customer problems is a key differentiator in shaping positive post-adoption behaviors.

This study employed a quantitative research approach to empirically test the extended Expectation-Confirmation Model (ECM) in the context of AI-powered services in e-retailing. Data were collected from 542 Vietnamese e-retail customers using a snowball non-probability sampling method. Partial Least Squares Structural Equation Modeling (PLS-SEM) was utilized to validate the research model and examine the proposed relationships between the constructs. The findings revealed a significant sequential mechanism: AI-powered service quality positively influences 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 (O = 0.371) and perceived usefulness (O = 0.279) were identified as the strongest predictors of continuance usage intention (p < 0.001 for both). Customer experience was found to play a central mediating role in this process. The study's results provide empirical support for the extended ECM framework, demonstrating the interplay between AI service performance, multi-dimensional customer experience, and post-adoption behaviors in an emerging digital economy.

This study successfully extends the Expectation-Confirmation Model (ECM) by integrating a holistic evaluation of AI-powered customer service, incorporating multi-dimensional customer experience (hedonic and recognition aspects), and distinguishing perceived problem-solving ability as a precise functional assessment. These additions provide a more comprehensive explanation of how customers form post-adoption judgments in AI-mediated retail environments, particularly within an emerging market context like Vietnam. The findings offer practical guidance for e-retailers, demonstrating that sustained usage of AI services depends on investments in both technical performance and the quality of experiential interactions. By clarifying how AI service performance shapes customer experience, satisfaction, and continuance usage intention, this research provides actionable insights for designing more effective AI-powered retail strategies. Future research could explore the cross-cultural generalizability of these findings in other emerging economies, investigate the impact of different types of AI services (e.g., proactive vs. reactive AI) on customer experience, and delve deeper into the specific hedonic and recognition elements that most strongly influence customer engagement. Additionally, examining the role of trust and perceived risk in the adoption and continued use of AI-powered services would offer further valuable insights.