The task of product quality control in model risk management processes

Problem statement. Currently, the task of product quality control at the operational stage is becoming urgent, while the transition to business processes based on AI and ML is associated with the risk of financial losses due to errors in the operation of models. These losses can be comparable to credit, market or operational risks, which makes model risk management relevant. The purpose of the study is to set the task of product quality control in the processes of model risk management and to describe a specialized information system that simulates production processes in digital form and represents a digital twin of technological processes that uses enterprise information for modeling and rapid response to changes. Results. It has been established that in order to increase the efficiency of production preparation, end-to-end management tools are needed that contain information about all stages of the production process, simulate decision-making processes and automatically recalculate decisions when external or internal factors change. Practical significance. The proposed approach has been tested in the processes of managing the Sber model risk, individual elements of this approach have been tested in companies of the food and metallurgical industries.