Abstract. Nowadays, while the number of users of interactive software steadily increase, new applications and systems appear and provide further complexity. An example of such systems is represented by multi-device applications, where the user can interact with the system through different platforms. However, providing end-users with real capabilities to author user interfaces is still a problematic issue, which is beyond the ability of most end-users today. In this paper, we present an approach intended to enable users to modify Web interfaces easily, considering implicit user intents inferred from example interface modifications carried out by the user. We discuss the design issues involved in the implementation of such an intelligent approach, also reporting on some experimental results obtained from a user test.