Have you ever wondered how much is really true about the hype around workflow management systems (WFMSs)? Based on over three years of experience with implementing and running a WFMS in a large bank, this contribution represents an exceedingly well-crafted, highly constructive demystification of this realm.

The bank separated front-office work from back-office work in 2003, when creating complex product bundles for customers, called customer packages. While front-office workers would consult with the customer on a particular customer package, the back-office workers would be responsible for creating the sometimes vast array of individual products in the various application systems.

The paper, eloquently written and even with some suspense, describes the initially Microsoft Word-based communication between front office and back office, and how, subsequently, the back-office tasks have been replaced, in an evolutionary style, by afull-fledged business process execution language (BPEL) based WFMS. Brahe and Schmidt manage to present a convincing narrative, rather than merely summarizing in a dry, academic, or jargon-laden way yet another WFMS project. During the unfolding of the story, the authors elegantly elaborate on the business drivers and successes achieved, such as the automation of approximately 80 to 90 percent of all back-office activities, increasing productivity by a factor of two to three. In addition, they also identify highly insightful sources of success and open issues, and do not hesitate to contradict prevailing opinions of the digerati, by highlighting deficiencies of the traditional user-analyst-developer system development cycle, or the stability and performance problems of the deployed WFMS system.

Finally, Brahe and Schmidt supply their realistic estimation of the potential of WFMS systems, which I fully share: we are still “far away from a situation where ordinary workers are able to define and compose their own local computational workflows.” However, the current slate of WFMS products and technology is indeed proficient in orchestrating production type processes involving the integration of several different information systems (IS), where production type processes are mature, well structured, and have few exceptions. In such cases, they point out, one is likely to reap the benefits of automation, integration, monitoring and controlling, and flexibility/adaptability of one’s processes.