Computer-Aided Design (CAD) plays a crucial role throughout a product's life cycle, especially regarding detecting design errors early on. To date, for complex engineered systems, the detection of design errors in a digital mock-up. It is a reactive and laborious activity that requires a complete CAD design before searching for applicable design rules in a set of large and unstructured design guidelines containing hundreds or thousands of design rules and making sure that the design complies with the applicable rules. This paper will illustrate how a context-aware cognitive design assistant, a seamless, ubiquitous and intelligent computer program, can sense relevant data of a CAD context to provide, without explicit user intervention, proactive and on-the-fly recommendations of relevant design rules, where relevancy depends on the designer's task. To illustrate the feasibility of the proposal, the cognitive design assistant will be integrated with the 3DExperience CAD software. The functional demonstrator shows that this proactive approach will not only reduce design iterations usually due to a-posteriori checks but also deal with any kind of design rules rather than a specific set of geometry and topology constraints verified by design checkers based on IF-ELSE instructions.