One major challenge for businesses today is the openning of their information systems (IS), making them accessible to their partners over the Internet. To this aim, they expose interfaces of their IS via Web wervice technology which consists in loosely coupled components using XML documents and commonly agreed standards to communicate. But this component set is heterogeneous and changing despite new technologies like Web services which assure a minimum level of interoperability. As such, the automation of Web service composition is a key enabler for rapid deployment of new composite services. Past work on the semantic Web provides several high level solutions to discover potential connections and interactions between heterogeneous Web services. However, few or none of these tackle implementations issues related to XML messages management at the syntactic level. This article focuses on adapting some XML database solutions, specifically XML schema matching, to leverage the semantic matching analysis by providing a data transformation that enables data integration of exchanged messages. Our approach requires no semantic addition to already existing industrial engines to execute the resulting compositions. The preliminary evaluation results, provided by an implementation in an industrial project, shows the effectiveness of the proposed solution.