For a while, training needs were identified by supervising teachers based on a personal effort using a short-term and punctual strategy. Programmed training could respond to current teachers’ needs, with practical and hands-on training that can be applied directly with students and organized on site. Despite supervising teachers' efforts to implement targeted training programs, there are discrepancies between student opinions and teaching objectives. This finding remains valid and stems from the fact that training needs have never emerged from the results of student evaluation of teaching (SET) survey. Through this paper, our aim is to start from the SET process to generate training needs for Training of the Trainers (ToT) programs for a better-quality teaching that is effective for both teachers and students involved in the curriculum. In fact, the process of the SET begins by collecting students’ opinion on AI teaching through many questions. Responses are related to different degrees of students’ satisfaction. After that, an analysis of results is done with different indicators such as a visualization of results by class, by course or by teacher. For each indicator, measures are used to determine if it is necessary to take decisions by proposing adequate training. The selection of an indicator with measures is considered as a scenario. Scenarios can be defined by combining more than one indicator. The execution of these scenarios leads to recommendations for future professor’s training. In this way, we can consider that we are aiming to achieve standards 9 and 10 by addressing standard 12.
