Automated Process (Re-)Design Maximilian Röglinger1,2[0000−0003−4743−4511] , Christopher van 1,2[0000−0002−4317−8592] Dun , Tobias Fehrer1[0000−0002−8798−5724] , Dominik A. 1,2[0000−0002−5218−6463] Fischer , Linda Moder1,2[0000−0003−2362−4839] , and Wolfgang Kratsch1,2[0000−0001−9815−0653] 1 FIM Research Center, Universities of Augsburg and Bayreuth, Germany 2 Project Group Business & Information Systems Engineering of the Fraunhofer FIT, Bayreuth, Germany {maximilian.roeglinger,christopher.vandun,tobias.fehrer,dominik.fischer, linda.moder,wolfgang.kratsch}@fim-rc.de 1 Introduction Business process management (BPM) continuously attracts academia and prac- tice, as it is known to drive organizational performance [7]. Especially process (re-)design entails significant economic value by introducing innovation, reducing costs, as well as improving quality, productivity, and customer experience [22]. Thus, it is considered an essential phase in the BPM lifecycle [15]. Today, organizations must overthink their business processes at an increas- ingly fast pace, consider continuously rising customer needs, create novel process- based value propositions, and engage in innovation to stay successful [7, 13, 15]. Technological developments are rapidly gaining momentum, processes are at drift, and ever more players enter the global market, resulting in the organiza- tional environment becoming more volatile, uncertain, complex, and ambiguous (VUCA) [5]. Even though this poses pressure on organizations, it also offers a wide range of opportunities. While automation is prevalent in other BPM lifecycle phases (e.g., in process execution) [1], process (re-)design commonly requires manual activities such as traditional creativity techniques [15, 22], making it time-consuming and labor- intensive. Thus, automated process (re-)design holds high yet unexploited poten- tial for long-term corporate success since it could accelerate process (re-) design and make it more efficient as well as less dependent on human creativity. 2 Research Problem 2.1 Problem Description Along the BPM lifecycle, many data-driven methods have recently emerged. Enabled by the increasing volume of data, process mining techniques have been developed to identify and discover process models based on process logs [2]. Predictive and prescriptive process monitoring techniques nowadays allow for Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). 2 Röglinger et al. Automated Automated Improvement Automated Innovation process Incrementally improve an existing Radically create a new process (re-)design process automatically automatically Manual Manual Improvement Manual Innovation process Incrementally improve an existing Radically create a new process (re-)design process manually manually Incremental Radical process improvement process innovation Table 1. Process (re-)design matrix acquiring real-time insights into future behavior and results of running process instances and provide recommendations for optimizing process control [21]. Driven by the recent “hyperautomation” trend [19] and the widespread adop- tion of process-aware information systems, organizations increasingly aspire to leverage automation potential in the context of process operations [7]. Whereas process mining and monitoring primarily focus on (partially) automated process control, robotic process automation (RPA) has become the new “technological star” for the lightweight automation of process execution [20]. Although some research obstacles still need to be overcome, ever more organizations adopt RPA to reduce manual efforts when performing specific tasks in processes [20]. Despite all these automation efforts, it is remarkable that the BPM lifecycle phase process (re-)design remains a manual task with a high level of cognitive effort. To illustrate the level of automation in the context of process (re-)design, we propose a 2x2 matrix along two continua (Table 1). The first continuum con- cerns the degree of automation (manual to automated process (re-)design), the second concerns the scope of process (re-)design (incremental process improve- ment to radical process innovation). In the following, we describe the state-of-art of process (re-)design within the introduced quadrants. With a lens on incremental process improvement, various collections of pro- cess redesign patterns and methods have been developed [8,12]. These collections reduce the cognitive effort and guide process stakeholders in process improve- ment. However, they neither replace manual effort nor do they leverage the potential of tools in the redesign process. Initial approaches for semi-automated process improvement have been developed (see [3]). Yet, these methods are at the lower end of automation, as they generally guide improving processes in a user-interactive way. Thus, there is still great potential to increase the level of automation. Research is already striving to further automate process improve- ment, enabling automatic exploration of beneficial process changes [9]. Focusing on radical process innovation, efforts also have been made to de- velop guidance for creating new processes with new value propositions [13]. For instance, Grisold et al. [14] created the “Five Diamond” method, which aims to guide organizations in identifying opportunities from business and technol- Automated Process (Re-)Design 3 ogy trends and integrating them into processes with novel value propositions. Nonetheless, equivalent to manual improvement, the introduced method does not support replacing manual efforts with automation. While automating pro- cess improvement seems easier to realize, the automation of process innovation proves to be an unsolvable problem to date. There is certainly still huge potential in the area of automated innovation that has barely been exploited. Overall, we conclude that process (re-)design is still predominantly a manual task. The automation of process (re-)design, especially with a focus on process innovation, undoubtedly remains a major hurdle to overcome. 2.2 Challenges to Overcome Several characteristics of business processes and the complexity of the process (re-)design task itself make the BPM lifecycle phase of process (re-)design stand out and, therefore, prevent or at least complicate its automation. Such charac- teristics are described here in broad strokes: Process (re-)design requires creativity. Process (re-)design often requires breaking with existing structures and routines within the process to create some- thing new. Falling back on existing concepts might be a good idea for evolu- tionary process improvements. Still, radical (re-)design relies on going beyond what has already been there and exploring the whole solution space of (possibly unknown) process (re-)design opportunities. In contrast to data-based improve- ment, such explorative and innovative efforts mostly rely on “creativity”, i.e., the use of imagination or original ideas to create something new. Creativity is often described as an inherently human capability. Therefore, automating (re-)design efforts requires advances in computational creativity. Processes are multi-dimensional. (Re-)designing processes is not as straight- forward as simply rearranging the sequence of activities within the investigated process. Business processes are commonly conceptualized using five fundamen- tal perspectives [23]. Besides the above-mentioned control-flow or behavioral perspective, these perspectives relate to the functional elements of a process (functional perspective), the assignment of tasks to human participants (organi- zational view), the implementation of atomic activities (operational perspective), and the information entities handled during individual tasks (informational per- spective). All perspectives have to be considered in automated (re-)design efforts. Processes are executed in context. Business processes are often part of an organization’s process landscape and, therefore, situated within a complex net- work of dependencies such as restricted resources, logical relationships [17], and domain-specific characteristics [4]. This makes it very hard to consider process (re-)design as a clearly delimited activity and, thus, complicates automation. Processes are socio-technical. Processes are sets of activities in which humans and technology co-create value [10]. Automated approaches in every phase of the BPM lifecycle are constricted by what data is available on these activities. Pro- cess mining can, e.g., only discover processes when their activities have left traces in the involved information systems or have otherwise been recorded [16] and when these traces are of high quality [4]. In return, this data represents only the 4 Röglinger et al. technical perspective on the process. Process and domain knowledge of human agents participating in or being responsible for the process is essential in guiding any (re-)design effort, making full automation impractical, if not unfeasible. Processes are at drift. All organizational concepts are subject to unintentional change, i.e., the deviation from their planned purpose over time. In a VUCA world, processes are no exception, constantly suffering from gradual and incre- mental changes over time called process or concept drift [6] or being radically changed by disruptive shocks [18]. This impacts automated process (re-)design activities twofold: First, due to such drift, processes are dynamic, constantly changing, and event-driven artifacts that are difficult to fully capture, define, and reinvent using high-level process models. Second, in a dynamic and changing en- vironment, attempting to (re-)design business processes is “subject to resistance, deals, side effects, and the properties of the IT landscape” [6, p. 193]. 3 Directions Towards a Solution Initial ideas towards a solution may involve approaches that leverage advances in computational creativity, e.g., evolutionary computation [3] or generative ma- chine learning [11]. Further, to accelerate process (re-)design, organizations could automatically incorporate feedback into design suggestions to shorten reaction cycles, e.g., via artificial intelligence-enabled process improvement tools and com- plex predictive models that capture trends from data. These initial ideas are beset with challenges themselves. For example, relying on historical data could lead to new processes already being outdated at the time of implementation. This demonstrates the need to address ancillary issues such as real-time data deployment [7]. A fully automated workflow environment would also be necessary to implement new process designs without delay. Additionally, artificial intelligence often works as a black box and lacks explainability. In conclusion, automated process (re-)design remains a relevant research gap that should be explored further. However, even if fully automated process (re- )design became feasible, new challenges would arise since organizations could then easily develop new processes. The focus of competition could move from talent to access to data or to the best forecasting models estimating the impact of changes on the future or identifying the next relevant time for re-evaluation. References 1. van der Aalst, W.M.P.: Business process management: A comprehensive survey. ISRN Softw Eng 2013, 1–37 (2013). https://doi.org/10.1155/2013/507984 2. van der Aalst, W.M.P.: Process mining: Data Science in Action. Springer, Berlin, Heidelberg (2016). https://doi.org/10.1007/978-3-662-49851-4 3. Afflerbach, P., Hohendorf, M., Manderscheid, J.: Design it like darwin - a value-based application of evolutionary algorithms for proper and unam- biguous business process redesign. Inf Syst Front 19(5), 1101–1121 (2017). https://doi.org/10.1007/s10796-016-9715-1 Automated Process (Re-)Design 5 4. Andrews, R., van Dun, C.G.J., Wynn, M.T., Kratsch, W., Röglinger, M.K.E., ter Hofstede, A.H.M.: Quality-informed semi-automated event log generation for process mining. Decis Support Syst 132 (2020). https://doi.org/10.1016/j.dss.2020.113265 5. Bennett, N., Lemoine, G.J.: What a difference a word makes: Understand- ing threats to performance in a vuca world. Bus Horiz 57(3), 311–317 (2014). https://doi.org/10.1016/j.bushor.2014.01.001 6. Beverungen, D.: Exploring the interplay of the design and emergence of busi- ness processes as organizational routines. Bus Inf Syst Eng 6(4), 191–202 (2014). https://doi.org/10.1007/s12599-014-0335-3 7. Beverungen, D., Buijs, J.C., Becker, J., Di Ciccio, C., van der Aalst, W.M.P., Bartelheimer, C., et al.: Seven paradoxes of business process manage- ment in a hyper-connected world. Bus Inf Syst Eng 63(2), 145–156 (2021). https://doi.org/10.1007/s12599-020-00646-z 8. vom Brocke, J., Baier, M.S., Schmiedel, T., Stelzl, K., Röglinger, M., Wehking, C.: Context-aware business process management: Method assessment and selection. Bus Inf Syst Eng (2021). https://doi.org/10.1007/s12599-021-00685-0 9. Dumas, M.: Process mining in 2021 and beyond (2020), https://www.linkedin. com/pulse/process-mining-2021-beyond-marlon-dumas 10. Dumas, M., La Rosa, M., Mendling, J., Reijers, H.A.: Fundamentals of Business Process Management. Springer (2018). https://doi.org/10.1007/978-3-662-56509-4 11. Elgammal, A., Liu, B., Elhoseiny, M., Mazzone, M.: CAN: Creative adversarial networks, generating art by learning about styles and deviating from style norms. arXiv:1706.07068v1 [cs.AI] (2017) 12. Fellmann, M., Koschmider, A., Laue, R., Schoknecht, A., Vetter, A.: Business pro- cess model patterns: state-of-the-art, research classification and taxonomy. Bus Process Manag J 25(5), 972–994 (2019). https://doi.org/10.1108/BPMJ-01-2018- 0021 13. Grisold, T., Gross, S., Röglinger, M., Stelzl, K., vom Brocke, J.: Exploring explo- rative bpm - setting the ground for future research. In: Business Process Man- agement, pp. 23–31. Springer, Cham (2019). https://doi.org/10.1007/978-3-030- 26619-6 4 14. Grisold, T., Groß, S., Stelzl, K., vom Brocke, J., Mendling, J., Röglinger, M., Rose- mann, M.: The five diamond method for explorative business process management. Bus Inf Syst Eng (2021) 15. Gross, S., Stelzl, K., Grisold, T., Mendling, J., Röglinger, M., vom Brocke, J.: The business process design space for exploring process redesign alternatives. Bus Process Manag J (2021). https://doi.org/10.1108/BPMJ-03-2020-0116 16. Kratsch, W., König, F., Röglinger, M.: Shedding light on blind spots: Developing a reference architecture to leverage video data for process mining. arXiv:2010.11289 [cs.CV] (2020) 17. Kratsch, W., Manderscheid, J., Reißner, D., Röglinger, M.: Data-driven pro- cess prioritization in process networks. Decis Support Syst 100, 27–40 (2017). https://doi.org/10.1016/j.dss.2017.02.011 18. Mendling, J., Pentland, B.T., Recker, J.: Building a complementary agenda for business process management and digital innovation. Eur J Inf Syst 29(3), 208– 219 (2020). https://doi.org/10.1080/0960085X.2020.1755207 19. Panetta, K.: Gartner top strategic technology trends for 2021 (2020), https://www.gartner.com/smarterwithgartner/ gartner-top-strategic-technology-trends-for-2021/ 6 Röglinger et al. 20. Reijers, H.A.: Business process management: The evolution of a discipline. Comput Ind 126 (2021). https://doi.org/10.1016/j.compind.2021.103404 21. Teinemaa, I., Dumas, M., La Rosa, M., Maggi, F.M.: Outcome-oriented predictive process monitoring: Review and benchmark. ACM Trans Knowl Discov Data 13(2), 1–57 (2019). https://doi.org/10.1145/3301300 22. Vanwersch, R.J., Vanderfeesten, I., Rietzschel, E., Reijers, H.A.: Improving busi- ness processes. In: Business Process Management. pp. 3–18. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-23063-4 1 23. Zeising, M., Schönig, S., Jablonski, S.: Towards a common plat- form for the support of routine and agile business processes. In: 10th IEEE International Conference on Collaborative Computing (2014). https://doi.org/10.4108/icst.collaboratecom.2014.257269