A deep analysis of the project's goals identifies the main need, whether that's to solve a task or automate processes through AI solutions. The result of this analysis makes building an expert system or neural network clearer. An expert system is man-made, a unique and very accurate model of the calculation is used. While a neural network is made using sophisticated machine learning. Engineers analyze what class the problem belongs to, the available data sets (quality and volume), and form signs from them. The network architecture is formed based on the type of task, and the learning process begins (architecture tuning, optimizer parameters, error function and form of signs).