Resmi Darni, Yulyanti Harisman, Lili Dasa Putri, Efi Fitriana, Rezki Ashriyana Sulistiobudi, Andhika Herayono
Phonological dyslexia is a neurodevelopmental learning disorder characterised by persistent difficulties in phoneme–grapheme mapping, phonological decoding, and pronunciation accuracy. Conventional dyslexia interventions are often non-adaptive, therapist-dependent, and limited in scalability, reducing their effectiveness in inclusive education contexts. This study aims to develop and evaluate a mobile adaptive multi-sensory artificial intelligence (AI) model integrated with real-time speech analytics to optimise phonological dyslexia intervention. The study employed a research and development (R&D) approach using the ADDIE framework, combined with a quasi-experimental pre-and post-test control group design. Participants were elementary school learners diagnosed with phonological dyslexia. The proposed system integrates multi-sensory phonological training, adaptive difficulty adjustment, and real-time speech analytics to provide immediate feedback and personalised learning pathways. Data were collected through expert validation, usability questionnaires, phonological reading tests, and speech analytics metrics, including phoneme error rate (PER) and word error rate (WER). © 2026 by the authors of this article.
Universitas Negeri Padang, Padang, Indonesia; Universitas Padjadjaran, Bandung, Indonesia
