The usage of workflow technology to model and execute business processes is widespread in many enterprises and within the software industry. With the growing maturity of sensors, wireless communication, and distributed computing environments, we can enhance this approach to enable smart workflows, which are business processes crossing the boundary to the physical world. Applications for such processes can be found in many application domains, like logistics or in smart factory environments. To realize smart workflows, workflow engines can be coupled with existing context provisioning systems. However, there is a gap between the rather low-level provisioning of context (e.g., sensor data and stock information) and the concepts needed in smart workflows (e.g., "is a spare tool available?"). The main contribution of this paper is to bridge this gap: we show how integration processes can be used to provide context information at different semantical levels for smart workflows.