COSTLEAP is an IRVA project, contributing to the Digital Production roadmap of Flanders Make. It is designed to optimize maintenance strategies by integrating cost-sensitivity in the learning of the data driven models and thereby also in the optimization of the moment of maintenance. The main goal is to shift maintenance decision-making from time-to-failure models to cost-centric models to allow companies to prioritize maintenance activities based on costs. To do so, COSTLEAP builds upon existing results to obtain a data driven cost estimation of maintenance (i.e. all costs related to (not) performing maintenance) by leveraging operational data mainly and primarily. Further, to improve the models, COSTLEAP will, in case available, leverage sensor data. Using these cost estimates it aims to enhance maintenance decisions both for companies with fixed and flexible maintenance intervals. To develop the data driven approach, next to public data sets, an emulator will be developed. Within the IRVA, also B2B projects with the companies of the user group will be defined to assess the feasibility of the full COSTLEAP approach for the specific company cases. Overall, COSTLEAP pursues a reduction of at least 10% of maintenance cost by designing cost-effective maintenance strategies aligned with operational goals of the companies.