This paper deals with artificial intelligence driven product engineering support. Many software systems are available to support the product lifecycle, especially during product design, such as CAD, PDM, CAE, SDM, etc. Most product development process is performed using these systems, which through their rich user interfaces allow skilled professionals to express their expertise and knowledge using the tools and functions the software is willing to provide them. At the end of the day, the result of their work is a model, built through a user interface, and stored in a repository. The goal of our research is to reverse engineer the user's knowledge by analysing his/her actions with the software system, based on the assumption that the process will itself be meta-knowledge driven and that we will focus on engineering software which provide semantically rich user interfaces. The aim of this paper is to investigate the idea of building reusa-ble expert knowledge from actions on engineering software user interfaces. It first outlines existing works from different fields and identifies remaining issues. It then suggest an approach to address these issues and put together an operational system.