Industrial AI for Maintenance in Sweden: Evolution and Industrial Challenges

Paper in proceeding, 2026

The growing convergence of Artificial Intelligence (AI) and industrial maintenance has led to significant advances in predictive and condition-based maintenance. This review paper examines the state-of-the-art in Industrial AI for maintenance within the Swedish landscape, focusing on how Industrial AI technologies are transforming traditional maintenance practices into proactive, data-driven frameworks. A comprehensive analysis of literature is presented, from the year 2000 until present, highlighting methodologies for fault diagnosis, predictive analytics, and decision support implemented within the Swedish manufacturing industry. Despite significant advances in computational power and data acquisition, large-scale industrial adoption remains limited. Data quality and availability, expert knowledge dependency, implementation cost, and organizational readiness persist as dominant constraints. Notably, the Swedish research landscape is characterized by strong academia–industry collaboration with a high proportion of real industrial case studies, providing valuable empirical grounding, while also revealing practical deployment challenges and offering unique lens to study how institutional factors shape AI adoption beyond algorithmic performance. The findings indicate that limitations to AI adoption in maintenance are increasingly structural and organizational rather than purely algorithmic. Future research should prioritize human-centered and hybrid AI approaches, standardized data management practices, explainability, scalable deployment models, and strategies to address data scarcity and trust in real maintenance environments beyond the pilot phase