The integration of Artificial Intelligence (AI) in predictive maintenance (PdM) poses challenges in ensuring user trust and understanding, especially for non-experts. This paper presents a visual analytics solution designed to support quick assessment of AI model quality and the need for retraining, using an interpretable machine learning (iML) approach. Developed with user-centered design principles, the visualization tool enables non-expert users to evaluate key metrics, including model performance and input data changes. The solution was applied to the maintenance of gas turbine blades and evaluated through focus groups with future users. Results show that visualization enhances understanding of AI operations and builds appropriate trust in AI systems. This work addresses a critical gap in Explainable AI (XAI) by focusing on practical, user-friendly tools for monitoring AI performance in industrial settings.