An integration model for adaptive learning in primary language education

This paper presents the experimentation of a digital and inclusive platform designed to support reading comprehension in primary school, with a focus on French as the language of literacy. Grounded in universal design for learning and adaptive learning principles, the platform integrates serious games, multimodal support tools (e.g. text-to-speech, visual lexicon), and AI-driven personalization to address classroom heterogeneity. Two experimental phases were conducted in French-speaking Switzerland with 150 and then 200 pupils, examining the impact of additional functionalities and adaptive learning systems. Data include usage logs, comprehension assessments, flow questionnaires, and teacher feedback. Preliminary results indicate that students actively engage with the accessible data, particularly when learning paths are personalized and aligned with their Zone of Proximal Development. The activation of adaptive learning seems to enable increased learner autonomy and motivation. Teacher feedback reveals strong acceptance and a perceived enhancement of differentiation practices. We reframe the traditional learning triangle into a ‘didactic pyramid’, positioning the digital system as an active mediator. We then propose an integration model based on two feedback loops — pedagogical and data-driven — that support sustainable implementation. Findings highlight the potential of adaptive CALL environments to foster inclusive, student-centred literacy instruction in primary education. Future work will focus on longitudinal learning outcomes and scalability.