Analyzing Learner–AI Interaction Patterns in Microlearning Environments

As artificial intelligence (AI) tutors become increasingly integrated into digital learning environments, understanding how learners interact with them is essential for evaluating their educational value. However, empirical evidence on actual learner–AI interaction patterns in authentic microlearning environments remains limited. This study examines learner–AI interactions in a microlearning environment using authentic usage data from a technical learning platform. A mixed-methods approach combines quantitative analysis of interaction and learning outcomes with qualitative analysis of tracked instructional support patterns. The analysis included 132 learners and more than eight hundred learner–AI exchanges. The findings show that interactions are primarily centered on clarification, explanation, and task-oriented guidance, highlighting the role of the AI tutor as an on-demand learning support tool. While the tutor effectively provides immediate explanations, reinforces learning, and sustains engagement, it demonstrates limited capacity to support deeper cognitive processes such as misconception diagnosis, multi-step reasoning, and reflective learning. These findings address the limited empirical evidence on how AI tutors are actually used and provide a clearer characterization of their instructional role within authentic microlearning environments. The study offers an empirical perspective on AI tutor behavior in microlearning environments and highlights directions for the development of more pedagogically intelligent learning support systems.