Adaptive reading assistance can improve the reading performance of students, but current dyslexia pedagogical theories do not yet provide sound results on a micro-level. We want to provide a reading assistance solution that both helps the learner and the dyslexia researcher. In order to archive this, we encode adaptation knowledge in a descriptive way by making use of state-ofthe-art ontology-based techniques. This enables a closed-loop approach of continuous improvement. In this paper, we want to present the overall approach as well as initial results of our work within the EU project AGENT-DYSL.