Modern data analysis in an industrial setting is not only concerned with the right choice of an analysis model, but to maybe even greater extent with integrating a model into an operational system and adapting it to a changing business environment. All of these steps are nowadays still being done manually by different experts. In this paper, we discuss approaches to automate these steps. We show a way to build an analysis model based on user preferences and introduce consistent concepts for the other steps. Due to the inherent fuzziness of the area, fuzzy systems turn out to be a natural choice for the underlying mechanisms