The typical 22MnB5 steel sheet metal used for hot stamping is precoated with an aluminum-silicon (Al-Si) coating in order to prevent surface oxidation and decarburization during the austenitization process, which occurs within a furnace. During austenitization, the Al-Si coating melts at ~577°C, followed by the transformation of liquid coating to Fe-Al and Fe-Al-Si intermetallic phases, which affect the as-formed performance of the coating when considering welding and corrosion. Understanding the transformation of these Fe-Al and Al-Fe-Si intermetallic layers is crucial to the optimization of the austenitization process, which strives to reduce furnace time. This research investigates the comprehensive effect of several heat treatment parameters including heating rate, soaking time, and soaking temperature on the intermetallic coating evolution of 22MnB5 sheet metal with two commercial coating weights 80 g/m2 (AS80) and 150 g/m2 (AS150). The study is divided into three parts. Part 1 of this study investigated the effect of heating rate and coating weight on the intermetallic coating growth under four different heating rates. SEM and EDS results revealed that the Fe2Al7Si (𝜏5) phase was first formed at the interface between coating and steel substrate, followed by the formation of FeAl3 (θ), Fe2Al5 (η), and a Fe-Al-Si ternary intermetallic precipitate at the boundaries of FeAl3 and Fe2Al5 for all the heating rate tests, which indicated that the heating rate did not change the species of the intermetallic phases, but it affected the transformation time. Part 2 of this study assessed the coating growth at different heating rates,soaking temperatures, and soaking times. It was found that the heating rate had more effect on the morphology of the thin Fe-Al-Si intermetallic phase compared to the soaking time and soaking temperature. The multilayered phase structure formed at heating stage was gradually transformed into FeAl (𝛽2) due to the interdiffusion of Fe and Al. The interdiffusion layer (IDL) consisted of FeAl, Fe3Al, and 𝛼-Fe. The thickness of the IDL increased for longer soaking time. When compared to the same soaking time, higher soaking temperature resulted in a thicker interdiffusion layer. The coating weights did not show a notable effect on the growth of the interdiffusion layer. Part 3 of this study was the development of an artificial neural network model to predict chemical composition of the transformed coating with different user inputs. The model was trained and optimized until the minimum Mean Squared Error (MSE) validation loss was achieved. Additional two stage heating tests were conducted, and the results were used as an additional validation to the neural network model. The model successfully predicted the training data and the validation test data as well.