Wiley’s learner-facing products already make use of recommender-based adaptive learning systems and knowledge state modeling techniques. With rapid advances in LLM technology, we have started renegotiating the traditional interfaces between our learning activities and our learners, which also requires a shift in design, development, and testing philosophy. In this talk, I’ll discuss the frameworks in which we’re pursuing problem-solving assistance and user-LLM-text interactions and demonstrate some work in progress.