The goal of the project was the development of software-supported procedures that "observe" users during the execution of a process with the software and write this information in fine granularity in an event stream or an event log. With the resulting online process analysis, dedicated processes are determined retrogradely from a running application and the use of the application is documented across different users, locations or points in time.
Scientific and technical starting point
The common approaches to obtaining models for business processes have in common that the software systems to be analyzed want to transfer as accurate an event stream as possible to a process mining tool. In 2016, the IEEE defined the XES (Extensible Event Stream) standard format for event streams. This standard is supported by common process mining tools and allows, among other things, free extensibility using custom extensions to map specifics of the company or its business processes, as well as specifics of the software systems and to transfer the event streams to the process mining tool. For the large standard systems with constant data models (e.g. SAP, Microsoft Dynamics, Oracle E-Business Suite), there are already preconfigured export routes that extract all the necessary event stream information from the application for a classic procurement process, for example.
However, there are no simple export routes for standard applications or industry solutions, let alone for individual software. Since in most cases the individual software solution does not necessarily use a workflow or process engine in the background that is capable of providing helpful audit data with just a few extensions, the only option in many cases is to look at the changes in the data model. However, as discussed above, this involves a lot of time wasted on interviews, analyses and assumptions.
As part of the research and development work, it became clear that the greatest innovation within the work packages lies in the identification of processes based on user interactions, identification of the "real" data model and the DataActivityLog. As long as a process description with BPMN or PHILharmonicFlows can be created from this, the further innovations in the area of conformity and compliance were only minor. For this reason, the focus in the further course of the project was much more on high quality in the semi-automatic recognition of the business processes implemented in the application.
The results and findings achieved in the project will be incorporated into the upcoming release of the PITSS.CON product and will therefore be available to the customer base as an update or for new customers as a stand-alone product.
DBIS supported PITSS in the project with the latest scientific and research findings and contributed its expertise in the field of business process management to the project. It also designed and evaluated the algorithms and procedures for determining the conformity and compliance of implemented processes with a given target process model or with existing compliance rules. DBIS has also published the particularly original results of its research work in the SoftProc project in prestigious publications and presented them at 12 international conferences and workshops.