This study aimed to examine how experts in AI-based learning perceive the categories of learning platform components to which AI technologies are applied, as well as the interrelationships among these components. It also conducted a theoretical and empirical analysis to derive a compositional model suitable for the ultimate purpose of learning. To achieve this purpose, a mixed-methods approach was employed, combining quantitative survey data with qualitative findings from expert focus group interviews. The analysis categorized functional components into two major contexts: a technological context, including AI algorithms, data analytics, and feedback systems, and a learning context, consisting of learning content, instructional strategies, and learner interaction. According to the results of the quantitative research, sub-factors within each context were categorized and analyzed to investigate whether they exist in a relationship with other sub-factors. As a result of this study, experts recognized the necessity of dynamic integration to support meaningful learning processes between technology and the learning context. Experts also hold different perceptions, depending on the group’s perspective on the interdependency among sub-factors of an AI-based learning platform. However, experts shared a common opinion that AI technologies must interact with educational goals and adapt to learner’s cognitive, behavioral, and affective characteristics in real-time, using data-driven insights generated by the learners. Accordingly, this study developed and proposed a model in which technological and pedagogical elements are functionally integrated within a data-driven context. The resulting learner-centered model envisions the future of AI-based adaptive learning platforms, offering both a theoretical foundation and practical guidance for system design.