Student models in educational software typically analyze how learners answer questions. But tutoring involves back-and-forth dialogue, which is harder to model. An arXiv preprint (2610.19532) argues that as tutoring becomes more accessible at scale, it is important to analyze dialogues between tutors and students as well.

The paper outlines a unified response framework that treats both question responses and tutoring dialogue within a single structure. The abstract does not provide implementation details, but the motivation is clear: current models miss the conversational nature of tutoring.

Since the source is only an abstract, we cannot confirm the method's specifics or evaluation. The claim is that a unified approach is necessary for large-scale tutoring.