This paper is concerned with a methodology and workflow developed to apply Artificial Neural Networks (ANN) to convert seismic attributes into a reservoir property. As a result, a Vshale cube was created for a complex fluvial-deltaic reservoir in the Azeri-Chirag-Gunashli field in the South Caspian Basin, offshore Azerbaijan. Selected well data was used as an input into the learning process of Multi-Layer Perceptron type of ANN technique to identify the relationships between seismic attributes and Gamma Ray log. The trained model was later applied to an attribute volume to cover a bigger area of the field away from the well control. Particular attention was given to preparation and quality assurance of the seismic data, which played a crucial role in improving the ANN training process. Four different testing techniques were used to validate the outcomes of the model using static and dynamic well data. Net-to-Gross maps, Vshale cross-sections and connectivity analysis from the well surveillance data were amongst the techniques used to verify ANN outputs both qualitatively and quantitatively.