Usable but Not Yet Fully Effective? Designing GenAI Learning Support Systems for Higher Education

Generative artificial intelligence (GenAI) is increasingly used in higher educa-tion, but its value for learning depends not only on usability, but on whether students can use it effectively to support meaningful learning-related action. This creates a central tension: GenAI can provide immediate and personalized support, yet learners may remain uncertain whether its outputs are reliable enough to act upon. Addressing this challenge, this study follows a Design Science Research approach grounded in the Theory of Effective Use (TEU). Based on literature analysis and student interviews, we derive meta-requirements and design principles for GenAI-enabled learning support and instantiate them in a web-based prototype. A formative evaluation with prospective users indicates excellent perceived usability and positive percep-tions of transparent interaction and informed action, but weaker representation-al fidelity. The study contributes TEU-grounded design knowledge for GenAI learning support systems and highlights actionable verification support as criti-cal for effective use.