Applications are a ubiquitous tool when running on mobile devices become an opportunity for a new digital business model. However, enterprises implemented the core business in legacy applications that are not designed to run properly on mobile devices. Consequently, this migration involves time and cost. The main contribution of this study to these limitations is an approach to assist this migration based on Markov chains that is a predictive user behaviour method that allows that available functionality is prioritized, the experience of User Interface is re-thought, and the new design is evaluated with a preliminary running version of the legacy app. An experiment on ten participants shows that a dynamic user interface that implements the Markov model provides the users with better usability, evidenced with an improvement of productivity labour and the Nielsen heuristics comparison between versions. This work suggests this model makes possible a smart and proactive interface design.