Process mining is the missing link between process science and data science. Process science is an umbrella term for the broader discipline that combines knowledge from information technology and knowledge from management sciences to improve and run operational processes. Business Process Management, Workflow Management, and Operations Research are subdisciplines of process science. Data science is a huge interdisciplinary field aiming to turn data into real value. Data may be structured or unstructured, big or small, static or streaming. The value may be provided in the form of predictions, automated decisions, models learned from data, or any type of data visualization delivering insights. Data science includes data extraction, data preparation, data exploration, data transformation, storage and retrieval, computing infrastructures, various types of mining and learning, presentation of explanations and predictions, and the exploitation of results taking into account ethical, social, legal, and business aspects.
The positioning of process mining as the connection between process science and data science, triggers the question how process mining relates to hyped terms such Machine Learning, Artificial Intelligence, Robotic Process Automation, and Hyperautomation.
Machine learning is concerned with the question of how to construct computer programs that automatically improve with experience. Machine learning is related to Artificial Intelligence. The field of Artificial Intelligence is broad and has different schools of thinking. Symbolic artificial intelligence uses high-level "symbolic," human-readable representations of problems related to logic and search. Symbolic AI was the dominant paradigm of Artificial Intelligence research from the mid-1950s until the late 1980s. Since the 1990-ties, the focus shifted to non-symbolic AI, for example, using artificial neural networks that are trained using large amounts of data. Neural networks tend to be huge and, unlike symbolic approaches, cannot be interpreted by humans. Deep learning using neural networks can be seen as a particular form of machine learning. Although the field of machine learning is much broader, progress in deep learning has been most visible. For example, speech recognition, as used by Alexa and Siri, build on this form of machine learning. This generated a hype where there is a lot of excitement around this particular form of machine learning and AI. However, the field of data science is much broader, and the successful applications of neural networks are limited to well-defined supervised learning tasks. Many organizations wasted money and applied the wrong technology to the wrong problem. Process mining is much more concrete and applicable in any organization. Process discovery and conformance checking have nothing to do with mainstream machine learning techniques like deep learning. Process discovery is closer to Petri-net-theory and pattern mining. Conformance checking is closer to optimization.
Although the core process mining approaches are different from machine learning, they can generate interesting machine learning problems. This is most relevant for the fourth type of process mining: operational support. Neural networks can utilize conformance-checking results to predict deviations, and use process-discovery results to predict the remaining flow time of a case.
Process mining is also related to novel forms of automation, such as Robotic Process Automation and Hyperautomation. Robotic Process Automation, often referred to as RPA, is an umbrella term for tools that operate on the user interface of other computer systems in the way a human would do. Unlike traditional workflow technology, the information system itself remains unchanged. I like to refer to RPA as the "poor man's workflow management solution." Because humans are replaced by software robots without changing the expensive information systems, RPA is attractive in many areas. Process mining can be used to detect where and how to use RPA. Moreover, after implementing RPA, process mining can be used to monitor the correct behavior of software robots. In fact, process mining can be used to oversee the cooperation between people, information systems, and robots. Therefore, I predict that in a few years, the process mining market will be much larger than the RPA market.
The term hyper automation refers to the combination of Robotic Process Automation (RPA), Machine Learning (ML), Artificial Intelligence (AI), and process mining. The goal is to automate more and more knowledge work using novel forms of automation and learning.
Hyped terms such as Artificial Intelligence and hyper automation are overselling ideas that have been around for quite a while and may be misleading because "there is no such thing as a free lunch." Often solutions only work for particular tasks or require a lot of engineering. Process mining is more down to earth. Information systems are loaded with event data that are not being used yet. At the same time, there are many opportunities to improve operational processes.