Рет қаралды 3,766
This short video was based on presentations in the 2021 BPM Expert Forum and the 2021 ISSIP Discovery Summit. The talk centers around the question "Why did #BPM fail?". About a decade ago, #BusinessProcessManagement (BPM) had a negative reputation and was often linked to business process modeling (e.g., creating BPMN wallpaper) and workflow automation using #WFM of #BPM systems. Many considered #BPM to be dead, and it was clear that traditional business process modeling provides little value and traditional workflow automation often failed. Organizations associated BPM with making process models rather than diagnosing and improving processes.
There were three main reasons for this skepticism and disappointment. Old-school BPM tends to be (1) unable to capture human behavior, (2) unable to deal with the complexity of real-life systems, and (3) unable to realize actual improvements. Moreover, applying WFM/BPM technology was rather expensive. As a result, processes are still hardcoded in application software or not supported at all. Many processes also use software from different vendors, making seamless integration difficult and time-consuming. Although the "M" in WFM and BPM refers to "Management", the focus used to be on modeling and automation rather than management. Traditional WFM/BPM systems fail to learn from the event data they collect. Real-life processes are more complex than people like to believe. The well-known 80-20 rule applies to processes, i.e., 80% of all cases are rather simple, but explain only 20% of the complexity of the process. The remaining 20% of cases tend to be neglected by software and management, but consume 80% of the resources of an organization.
In this talk, I argue that both #ProcessMining and #RoboticProcessAutomation (RPA) served as lifelines for traditional BPM. RPA shifts the boundary of cost-effective automation and can therefore be seen as "the poor man’s workflow management solution". Process mining complements RPA by identifying routine work and monitoring processes before and after the introduction of RPA.
RPA can be seen as a bottom-up activity, i.e., removing repetitive tasks. Process mining can help to identify and automatically learn such tasks. However, the primary use case of process mining is the top-down analysis of end-to-end processes. Process mining techniques use event data to show what people, machines, applications, and organizations are really doing. Process mining provides novel insights that can be used to identify and address performance and compliance problems. Just like spreadsheets can do anything with numbers, process mining can do anything with event data, i.e., it is a generic, domain-independent technology to improve processes. Using process mining, one can uncover compliance and performance problems. It enables conversations anchored in reality instead of anecdotes, beliefs, and company politics.
The goal of the video is to trigger discussions on the role of BPM and the importance of #ProcessMining and #RoboticProcessAutomation.
Disclaimer: BPM can be interpreted in different ways. Half of the papers at today's scientific BPM conferences are about process mining and other data-driven approaches. When talking about "old-school BPM" this is not included. Process mining helped BPM to become relevant again (just look at the 40 process mining tools, the success of Celonis, and the takeovers of Signavio, Minit, myInvenio, etc. by SAP, Microsoft, IBM, etc.). Of course, process mining is one of many BPM techniques and does not replace (low-code) automation and modeling. However, it definitely helped to make the BPM community and vendors become more data-driven without dropping process-centricity altogether (like most BI tools).