Reference architecture and adoption roadmap for a manufacturer–end-user predictive maintenance ecosystem, a human-centered blueprint for manufacturing SMEs

Although predictive maintenance (PdM) is widely recognized for its strategic value, its adoption remains uneven across the manufacturing sector, particularly among manufacturing small and medium-sized enterprises (SMEs). These firms face persistent barriers to adoption, from unclear ROI to legacy-system integration and cybersecurity constraints. This research develops and exploratorily assesses an OEM-end-user collaborative PdM ecosystem using a Design Science Research methodology. The artifact is grounded in a systematic review of 123 studies spanning adoption barriers, contribution frameworks, and adjacent maintenance. The study’s contribution lies not in a new PdM algorithm, but in integrating five elements usually treated separately socio-technical framing, a configurable reference architecture, a progressive roadmap, a trust-by-design pack and OEM-end-user value exchange into a single ecosystem, with AI positioned as an an enabler, not a prescription. The proposed artifact was assessed through three rounds of modular expert surveys with non-overlapping participants. Findings confirm PdM adoption as fundamentally a socio-technical challenge rather than a technical one, with experts endorsing progressive deployment, human-in-the-loop oversight, efficient data use, and governed OEM-end-user collaboration as the conditions that make adoption work. This study offers an evidence-informed, expert assessed blueprint bridging PdM technical maturity and industrial adoption, laying the groundwork for future field evaluation.