The study examines the knowledge representation models developed for the CleverCOMSRL intelligent expert system in terms of their ability to resolve inconsistencies. Four knowledge representation models (KRMs) are considered: production, semantic network, frame, formal logic. The main directed activity is prediction, forecasting within a certain aspect of the subject area. The use of expert systems makes it possible to merge the design of new products and the technological preparation of their production into a single process, which contributes to the full disclosure of human creativity. An automated expert system is considered, which enables the automatic construction of the mathematical model of the CleverCOMSRL system when the designer specifies the composition of its units and the scheme of their connection. Variation modules <Space Path=""> of the CleverCOMSRL system are developed and experimentally tested. An algorithm is proposed to search for internal contradictions by determining all satisfying conditions while checking the efficiency of an intelligent expert system. A functional is developed considering the structure necessary to perform intelligent interaction in the context of performing behavioural analysis of an intelligent agent. Components of architectural compositions have been designed taking into account shapes and volumes, through intelligent lighting control and the possibility of selecting functional criteria of variant models of the CleverCOMSRL system.</Space>