Operation of any type of industrial or commercial facility depends on the use of electrical energy. To that end, facilities are connected to the power grid in order to have access to continuous electrical power supply. Therefore, analysis of power consumption at the grid's nodal level may provide information about the consumers connected to the node. In this paper, a new intelligent system for detecting hidden facilities by utilizing consumption signals measured in smart grids is presented. In particular an anticipatory system that makes use of the synergism of Gaussian process regression and fuzzy logic is adopted for inferring the presence or not of hidden facilities. Presented scenario demonstrates the potentiality of the method for accurately detecting anomalies in power consumption that correspond to hidden facilities.