The majority of web applications nowadays are data-driven. However, that does not mean that all data is available while launching the respective application. While considering Web 2.0 applications, data is often fetched on-demand from remote web services, for example, after a location was provided, weather data could be fetched, and local news is loaded. This mashup approach is highly dynamical, i.e., based on the data input of the user, completely different execution paths might be performed. Currently, such workflows are implemented within the application logic requiring high development effort and maintenance of the implemented logic to prevent unintentional behavior. In this paper, we present a novel approach to integrate web services dynamically, to decrease deployment and maintenance costs and to enable the next generation of interlinked data web applications to enables application architects to (re)define the data integration in a descriptive way in an ontology, validate the workflows and define logical requirements. However, our approach is not just a design method but also a method for ad hoc integration of new services. Our approach has a significant impact on the effort for generating and maintaining dynamic applications.