Wharton-backed research on AI tutors reinforces a key idea for Edzy: the real moat is not the chatbot, but the engine that decides what each student should practice next.
A useful insight from Wharton’s coverage of AI tutoring is that the real leap does not come from just answering student questions. It comes from deciding what the student should practice next based on how that student is actually performing.
The research it highlighted, from a five-month Python course across 10 Taipei high schools, showed that students receiving AI-personalized problem sequences outperformed students following a fixed easy-to-hard sequence, without needing extra teacher workload or extra instruction time. That is a strong signal for where learning products should pay attention.
For Edzy, this reinforces an important product belief: the moat is not just the chatbot. The moat is the personalization engine that uses student attempts, doubts, time spent, mistakes, and interactions to determine the next best question, hint, revision, or concept. Adaptive practice, mastery scores, productive struggle, and a dynamic learning path for every student feel like the bigger opportunity.