This paper presents the research work towards improving human computer interaction by providing intelligent assistance to users. This has been approached by incorporating principles of a cognitive theory in a Graphical User Interface (GUI), that deals with file manipulation and is called IFM. The cognitive theory is called Human Plausible Reasoning (HPR) and has been used to simulate users' reasoning in the user model of the system so that the GUI may provide spontaneous assistance to users' errors. Such a goal is difficult to achieve and depends heavily on the development process. However, there is a shortage of reports on the software engineering process of intelligent assistants. Moreover, in the literature of intelligent assistants there is evidence that some important phases of their development process may have been omitted and thus the understanding of delicate issues has not improved significantly. Therefore, the focus of this paper is on presenting and discussing the software engineering process of the intelligent assistant developed. Special emphasis has been put on the description of the experimental studies, which were conducted prior and after the development of the system. Theses studies were used for the specification and refinement of the overall design as well as the adaptation of HPR in it. The experimental results have shown that the intelligent assistant may follow the users' reasoning and provide helpful advice to a satisfactory extent as compared to human advisors.