BELIEF - From Prediction to Trust: AI-Based Decision Support for Maintenance

Purpose and goal

BELIEF develops AI-based decision support for predictive maintenance in Swedish industry. The project combines uncertainty-aware machine learning, explainable AI, and rule-based reasoning to provide reliable and understandable recommendations. Through close collaboration with industry partners, solutions are developed and validated to strengthen trust in AI, reduce unplanned downtime, and improve resource utilization.

Expected effects and result

The project is expected to deliver AI-based decision support solutions for predictive maintenance that have been validated in industrial environments. The results will contribute to reduced unplanned downtime, more efficient resource utilization, and increased operational resilience. The project will also strengthen companies’ ability to develop, adopt, and use trustworthy and explainable AI through new methods, demonstrators, and frameworks for AI maturity and organizational development.

Planned approach and implementation

The project is carried out in close collaboration between academia and industry. Through co-creation, key maintenance decision situations are identified, after which AI-based methods for uncertainty handling, explainability, and rule-based reasoning are developed and integrated into decision-support solutions. The solutions are validated iteratively in industrial environments, with a focus on usability, trust, and organizational adoption.