AI-Enhanced Gamified Nudge?: Supporting Self-Regulated Learning in Higher Education: A Design Science Research Study

Nikodem and I wrote the thesis as equal co-authors. I led the research from end to end: the theory, the study protocol, the instruments, the analysis and the writing. I also designed the flows, wireframes and interface of Regula, the platform we built for the study. Nikodem worked on the research design, the interface, the prototype and the writing, and helped with the analysis code in R.

AI can plan a student’s week for them, and that’s the risk: planning, monitoring and reflecting are the skills self-regulated learning depends on. We asked what happens when AI is held to a few narrow jobs inside a gamified learning platform.AI can plan a student’s week for them, and that’s the risk: planning, monitoring and reflecting are the skills self-regulated learning depends on. We asked what happens when AI is held to a few narrow jobs inside a gamified learning platform.

Following Design Science Research, we designed and built Regula, where AI does three things: it breaks assignments into daily tasks, comments on submitted drafts and recommends reading. A second version left those three features out. In a two-week quasi-experiment in April 2026, 27 students were randomly allocated to one version or the other. We combined system logs with two validated questionnaires, the MAI and the MSLQ, analysed them with non-parametric tests in R, and ran a thematic analysis of the students’ written responses.

AI support raised the number of tasks students created, with a large effect (p = .028), but completion rates didn’t differ between the groups: planning went up and doing stayed where it was. The Controlled group wrote more than twice as many reflections. AI-enhanced participants also rated their own self-regulation higher than their logs suggested.

That last result came from ten questionnaire responses and didn’t hold once the study grew to 54 students. The Regula case study has the full-sample findings.