This paper presents an analysis on the value of condition monitoring considering the uncertainties inherent in RUL (Remaining Useful Life) prognosis, the risks entailed by prognostic inaccuracies, and the essential investment prerequisites associated to the implementation of modern digital maintenance approaches. In particularly, we explore this in the context of fleet maintenance using an optimisation model that integrates predictive maintenance with existing preventive maintenance and operational schedules workload balance dynamically. The model presents a novel approach, combining the aforementioned dimensions with a holistic approach, that corresponds to an unsolved gap in literature. The primary objectives of the model application are to ascertain the conditions under which introducing predictive maintenance methodologies generates value, and to identify the junctures where the investment necessary for model implementation becomes economically justifiable. Furthermore, the paper will discuss the long-term benefits that materialise over an extended time horizon. The analysis shows how different levels of RUL prognosis uncertainty affect the fleet allocation to maintenance and operation, and the cost savings that reducing uncertainty means for fleet management. The costs are calculated executing the model in two exemplar cases using Python and Gurobi solver. Based on the savings comparing two uncertainty cases for RUL, companies can estimate the required time to amortise their investment in condition monitoring.