As large language models make on-demand tutoring technically feasible at scale, higher education faces a central question: can AI-based tutoring meaningfully improve learning outcomes when embedded in real university courses rather than used as a generic study aid?

One such system has been adopted by more than 27,000 students across 40 higher-education institutions and over 900 courses, providing lecture-specific AI tutoring and quiz-based practice integrated into existing teaching structures. Empirical evidence from a large introductory computer science course at the Technical University of Munich indicates that sustained use of AI-supported tutoring and assessment is associated with a 21% reduction in failure rates. These findings highlight the potential of course-integrated AI tutoring to contribute to improved learning outcomes at scale.

Prof. Dr. Alexander Pretschner

Professor at Technical University of Munich

Philipp Csistian

CEO at OneTutor GmbH

Philipp Csistian initially studied Business Administration at the Technical University of Munich (TUM) and gained his first professional experience in the field of Private Equity. His bachelor's thesis on a digital matching platform sparked his interest in Computer Science, a subject in which he subsequently completed his Master's degree, also at TUM. During his studies, he worked at Amazon and AWS at the intersection of technology and business. Philipp Csistian is CEO & co-founder of OneTutor.