This paper focuses on the design and implementation of an AI-based architecture diagram. Each module of this diagram contributes to the generation and programming of different components, namely robot, fixtures, grippers, fences etc. that are necessary to create a robotic cell in virtual environments. Additionally, the application includes modules that perform self-correcting suggestions, checking for any anomalies, such as overlapping parts, in the generated scene, providing corrective feedback to the user. The aim of the architecture is to help with the creation of an AI-based application that would enable the users create demos of robotic cells in VR-based environments, similar to the real ones. In this paper, the background of the architecture is introduced, then a description of the architecture is provided. In the use case, an AI-based application is demonstrated as example, while in the end, results and next steps are enclosed in the conclusions.