Inclusive education aims to provide equitable learning opportunities for all students, including those with special needs, yet its implementation faces challenges such as limited resources and curriculum adaptation. Artificial Intelligence (AI) emerges as a promising solution to personalize learning and enhance accessibility. This study investigates the optimal utilization of AI in inclusive education management, focusing on strategies to support students with special needs while addressing implementation barriers. Employing a qualitative case study design, the present research was conducted in AI-implementing inclusive schools in Mataram City, Indonesia. Data were collected through in-depth interviews with 12 participants, including principals, education managers, teachers, students with special needs, and parents, supplemented by participatory observations and document analysis. Thematic analysis was utilized for data interpretation, with triangulation and member checking ensuring validity. Key findings reveal AI's significant role in personalizing learning and enhancing accessibility through adaptive systems, speech-to-text, and text-to-speech functionalities, leading to improved student engagement and comprehension. However, implementation faces substantial challenges, including inadequate digital infrastructure, insufficient teacher training, and a lack of clear ethical and regulatory policies. To optimize AI's use, the study proposes four key strategies: strengthening teacher AI training, developing robust digital infrastructure, establishing clear AI policies and regulations, and fostering the development of more adaptive, human-centered AI. These findings offer theoretical implications for technology integration and socio-constructivist learning, and practical implications for policymakers, educational leaders, teachers, and AI developers in fostering truly inclusive and equitable educational environments.