The Future of Education

The integration of Artificial Intelligence in Higher Education is often limited to standalone tools, neglecting the structural complexity of Learning Management Systems (LMS). This paper introduces "Agent EDDIE" (Educational Didactic Digital Interchange Engine), a collaborative project between Mid Sweden University and Offenburg University of Applied Sciences. EDDIE is designed as an AI-agnostic agent capable of bridging the gap between LMS environments like Moodle and Canvas. Unlike provider-specific solutions, our architecture ensures strategic independence by decoupling the agent’s logic from any specific Large Language Model (LLM). At the current stage of development, the agent has successfully demonstrated core functionalities: it autonomously authenticates within Moodle, validates course accessibility via unique IDs, generates a comprehensive inventory of student-visible elements, and utilizes LLM APIs to synthesize rudimentary course descriptions. However, the primary focus of this contribution is the strategic roadmap of the project. We outline the planned evolution of EDDIE through four critical milestones: first, the activity-by-activity technical transfer between Moodle and Canvas; second, the integration of automated linguistic translation; third, the integration of Open Educational Resources (OER) from external repositories to enhance existing courses, and finally, the implementation of a didactic audit tool. This tool will evaluate courses against abstract "studyability" rubrics, providing instructors with feedback based on pedagogical best practices. By presenting this work-in-progress, we aim to discuss the challenges of maintaining identical usability across systems and the importance of academic sovereignty in the age of AI.