In shale reservoirs, quantification of total porosity and water saturation are considered key parameters for calculating the volume of hydrocarbon in place. However, the petrophysical interpretation necessary for the definition of these parameters represents a challenge due to mineralogical complexity of the reservoir as Well as the presence of kerogen and different fluids depending on their thermal maturity. Therefore, for a good petrophysical characterization, it is necessary to drill pilot vertical Wells to acquire a wide variety of data, both from logging tools and core samples. As a result of this, there is a rise in the budget associated with the acquisition of data in the shale characterization projects. On the other hand, there are many old Wells that have been drilled and logged in the Vaca Muerta-Quintuco formation section, due to their position in the lithological column overlying deeper conventional targets. The purpose of this work is to present a workflow that allows the use of data from these legacy Wells with a limited set of logs applying calibrations carried out using machine learning techniques based on probabilistic petrophysical interpretation carried out in Wells with a complete set of logs and laboratory data. Once the target log has been defined, an exploratory analysis of the logs available in the aforementioned set of Wells is carried out. The logs are studied separately and their relations as a whole, obtaining as a result a set of data with validated quality and with a format suitable for feeding machine learning models. Data is divided into training and validation sets. Then, different models are trained and validated, adjusting them to finally obtain the best result. The implementation of this technique not only provides relevant information for improving property maps and volumetric calculations, taking into account the uncertainty associated with the results obtained, but also allows to analyze variability of porosity related to other geological characteristics, such as the depositional system and the thermal maturity of the rock. Finally, a better use of the available data allows to reduce costs allocated to the acquisition of new data in areas where the information is available from previously drilled Wells.