Artificial intelligence (AI) is increasingly present in our lives, and it is generating a higher demand for human resources with knowledge about this area of computing. It is essential to have tools or resources that help in learning AI. This article presents an expert system that supports teaching the immersed process (workflow) in developing systems that use machine learning, particularly classification. This system is based on rules and preconditions guiding the development process, validating that the functions are executed logically and indicating errors made. The system architecture allows users to add or remove functions and precondition rules. One key role of the system is to guide users in identifying and solving errors. The qualitative evaluation showed that the students, after using the system for a few weeks, were able to confidently identify and solve errors when implementing practices that involved this type of development process.