The paper reviews the problem of human-computer interaction between engineers and their intelligent digital assistants in the process of computer-aided design. Affordance theory is used as a methodological tool for analyzing and monitoring this interaction. Based on the semiotic theory of Charles Sanders Peirce we introduce further definitions of similar properties and differences between the roles of human and Artificial Intelligence towards the designation process. Affordance can be introduced by the typology «Legisign-Index-Decisign» and defined as an apperception of the use case. As part of the study, an original experiment was conducted, which consisted of automated visualization of a sequence of 100 images of various technical devices, mostly antique and obsolete to eight people with different professional skillset. The subjects were asked to guess the purpose of each device, and as a result, the correctness of the answer and the heat map of the subject’s eye movements during the experiment were recorded. In general, in all cases, we observe more active visual activity and better result of the engineer, which indicates the skill of affordances recognition. Based on the results of experimental studies and theoretical reasoning, we clarified definition of affordance in AI supported engineering activity, introduced its role in engineering design as a general concept for human-machine interaction, and defined a pragmatic rule of affordance semiosis in predicative form. Visual activity of a human subject when indicating the purpose of a technical device is being associated with the search and recognition of affordances expressed in the form of corresponding signs. Artificial Intelligence in technical design plays a mediator role between the object and the subject, without replacing the latter. Within the framework of the considered topology of signs, Artificial Intelligence can promote creativity by offloading the routine job of searching for signs onto a digital assistant.