Abstract:

Integrating artificial intelligence into Virtual Learning Environments (VLEs) and Learning Management Systems (LMSs) requires more than adding new features. For higher education institutions, it involves balancing pedagogical value with transparency, data protection, governance, and long-term operational fit. This paper presents the experience of Sabrina Marquardt (Catholic University of Applied Sciences North Rhine-Westphalia) and Jesús Copado (SURLABS) in implementing SURContext in the open source LMS ILIAS and examines the conditions required to move from experimentation to institutionally viable adoption.

The study is based on a pilot carried out at the Catholic University of Applied Sciences NRW, where SURContext was introduced to provide AI-supported interaction within the university’s existing digital learning environment. Rather than treating AI as an external service, the project explored how context-aware, source-based AI can be embedded directly into the LMS in a way that aligns with academic workflows and institutional requirements. A central objective was to provide meaningful user support while ensuring that responses remained grounded in trusted materials and understandable in origin.

The pilot showed that trust, answer quality, and institutional acceptance were strongly influenced by source transparency, content governance, and the use of curated knowledge bases. Embedding AI within the LMS proved more effective when users could relate responses to institutional materials and when the system’s scope and limitations were clearly communicated. The implementation also confirmed that governance and communication are as important as technical performance in achieving responsible and manageable AI adoption in education.

The case demonstrates that context-based AI can be integrated into open source LMS platforms in a way that is educationally meaningful and operationally sustainable. It concludes that successful institutional adoption depends not only on technical integration, but also on the careful design of knowledge curation, transparency mechanisms, and governance processes.

Keywords:

AI, Agents, RAG, ILIAS, Moodle, LMS.