Industry 4.0 has driven significant transformation in manufacturing maintenance strategies, particularly in the implementation of Total Productive Maintenance (TPM). This research examines the paradigm shift from conventional TPM methods to Artificial Intelligence (AI)-based approaches, utilizing a qualitative methodology with phenomenological design and participatory observation in manufacturing environments. Results indicate that integrating AI with TPM has transformed maintenance approaches from reactive to predictive, enabling real-time anomaly detection, predictive maintenance, and maintenance schedule optimization that significantly improve Overall Equipment Effectiveness (OEE). Despite providing benefits such as reduced downtime and maintenance costs, AI-based TPM implementation faces challenges including data fragmentation, skilled human resource requirements, and organizational culture resistance. The development of collaboration between industry and academic institutions, as well as investment in edge computing infrastructure, are identified as key to maximizing the potential of AI-based TPM in enhancing operational sustainability in manufacturing.
