Process mining without contextual models provides only half the truth.
Like a detective, you have to analyze lots of clues in order to determine the entire picture.
The topic of process mining is enjoying an immense boost—not just due to digitalization, but also as a result of classic process topics such as increasing efficiency and reducing costs. However, the pure measurement and reconstruction of processes is only the first step to understanding and improving business processes. The true potential of process mining can develop only through the suitable linkage of this knowledge with additional contextual models.
The increasing digitalization of products and services compels businesses in all industries and of all sizes to reconsider their existing business models and the processes they implement. Here, more than purely technical challenges alone need to be mastered. In addition to well-known topics, such as standardizing internal processes to reduce costs while increasing speed, new challenges, such as the optimization of customer journeys to increase customer satisfaction, are increasingly gaining prominence. The business process management department is back in demand!
Process mining is certainly not a new topic; it is striking that process mining was previously used primarily by the operational technical side—less so by the BPM department. Today, BPM departments increasingly face the challenges of keeping the required process documentation up-to-date in an automatic way and recognizing problems of the actual implementation in time.
Here, process mining helps: Large quantities of data can be analyzed and concentrated into process knowledge. To this end, techniques in data mining are applied on event data and intelligent algorithms to generate process structures from process event logs. As previously indicated, process mining is a procedure that is driven primarily by data. It generates process structures and metrics that, like clues, are evaluated by means of analytical procedures and consolidated into Key Performance Indicators (KPIs).
These KPIs in turn are used in management dashboards as a basis for operational decision support. The results of process mining—structural process models as well as quantitative key figures—can be interpreted only with difficulty, though. Only in connection with process context and human expertise can technically valid conclusions be drawn.
It is to be assumed that artificial intelligence (AI) will be increasingly integrated in process mining. Thus, for example, distinctive features in process procedures can be automatically identified and suggestions for improvement offered. process mining tools will feature more of an assistance character rather than implementing fully automatic process changes. Furthermore, by using innovative deep-learning approaches, predictive models can be trained to allow for a reliable prognosis of process sequences and KPIs. This permits the implementation of proactive process control.
Also to be expected is the combination of process mining and Robotic Process Automation (RPA). Thus, many companies continue to apply legacy systems that make the use of process mining technology more difficult. RPA technology can understand the surfaces of legacy systems content-wise and in a data-technical way. This property can be used to provide data for process mining without intervening in legacy systems or needing to implement costly back-end extractions.
To conclude: Process mining generates actual process models and the associated KPIs from data clues. Without to-be models and corresponding context information, however, these are often only half the truth, since the benchmark, classification and assessment are missing.
For a beneficial application, the technology of process mining must be integrated into modern platforms of business process management, such as ARIS, and combined with methods of AI and RPA in the future.
Read the full article here.