Model-based reasoning can be considered central in very diverse domains of practice. Recently considered domains of practice are political discourse, social intercourse, language learning, archaeology, collaboration and conversation, and so forth. In this paper, we explore features of model-based reasoning in architectural design and construction. Additionally, an indication is given of some existing suggestions of how model-based reasoning systems may be simulated in an automated environment. We extend these lines of thought into our own simulated environment and give indications of how such a model-based reasoning system can not only give us better insights in the architectural design and construction practice, but also why it is so hard for such a system to eventually surpass human capabilities in this area of practice.