Providing certainty for investors: an industry standard for process mining

Commentary by Deloitte partner Olly Salzmann on the establishment of a reference model.

To create comparability and transparency in the process mining market, a reference model needs to be established as an industry standard. Such a standard would also provide the basis for innovations, for example to expand the scope of the system to include artificial intelligence, robotics or blockchain.

A lot has happened in the field of process mining over the last 15 years. What started out as a very academic discipline instigated by Dutch computer scientist Prof. Wil van der Aalst is now one of the fastest-growing technologies in the market. Adaptation of this technology at numerous Fortune 500 companies, the entry of SAP and other software giants, and Marc Kerremans’ recently published Gartner report have created a real hype.

Many professional service companies, from technology consultants to accounting firms to strategy consultancies, see great potential for the creation of their own dedicated discipline. The current discussions about digitalization, Industry 4.0, artificial intelligence, and big data are also accelerating the rise of process mining. There are now over 20 technology providers positioned in the three main areas of application - process modeling, process monitoring, and process discovery - with the number continuing to rise.
In essence, process mining supports companies in their digital transformation. Business and IT processes are made visible. When process mining is performed, all types of “analog” processes in place are recorded and thus made digitally usable. The transparency thus achieved can be capitalized on in many ways. For example, process efficiency can be increased, standardization expedited or compliance cases reproduced.

In spite of the considerable progress made in process mining in recent years, the industrial application and refinement of the technology is still in its infancy. Besides the integration of artificial intelligence (AI) for process simulation, rapid, efficient development of new data sources including all relevant company applications for the consideration of end-to-end processes is currently one of the biggest challenges.

As in the case of all promising but new technologies, a large number of users, consultants, and technology providers operate in an increasingly opaque market. To create comparability and transparency, it is imperative to establish a process mining reference model as an industry standard.

This industry standard will guide companies, technology developers and providers, and users in the validation, application, and further development of process mining. It will also provide certainty for investors, maximize the reusability of modules developed, and minimize dependencies on proprietary components.

The first promising research projects from industry, universities, and providers have already been initiated.
The aim is to develop a reference model that defines an industry standard which is accepted in the market. The standard consists of four stages that successively represent the lifecycle in the application of process mining:

1. Data provision (smart data discovery): identification and extraction of relevant information from source systems

In addition to the existing extraction tools for established IT systems, approaches will be developed and documented that automatically identify and extract the relevant information using AI. This will reduce the effort for data acquisition from individual workflow systems, in-house developments, and niche systems.

2. Data standardization (metamodel): standardization and harmonization of information from different source systems

System-specific designations should be converted into a uniform nomenclature and made comparable. This will simplify and accelerate the merging of additional data from heterogeneous system landscapes into a process mining framework.

3. Process modeling (process library): reconstruction of real process flows—digital process twins

Flexible process for linking process-related objects along the enterprise value streams, enriched by tax-related attributes. Established process modules will be combined in process libraries so as to continuously improve the application and the benefit (time-to-value).

4. Process visualization and analysis (analysis library): digital platform for the standardization of innovative business analyses

Development of industry- and process-specific analysis libraries with a clear focus on economic benefits. These templates will enable comparability between and within companies and create a digital platform for embedding innovative analytics. The simulation and the prediction of process flows are particularly important milestones along the road to automated recommendations for action.

The four levels of the reference model will enable companies to use the model efficiently and technology providers to develop it further for specific industries. At the same time, the industry standard will provide the basis for further innovations, such as for expanding the scope of the system to include artificial intelligence, robotics or blockchain, and then immediately industrializing it. 
The outlook for a sustainable, diverse application of process mining is extremely favorable and will facilitate the transformation from a straight process visualization into a comprehensive digital platform. In the coming years, the establishment of the process mining reference model as an industry standard will be instrumental in companies’ successful digital transformation.


Source: ManagerWISSEN, Harvard Business Manager, October/2018 issue

Der Industriestandard für Process Mining schafft Investitionssicherheit

open in new window >> Zum deutschen Artikel

Deloitte Center for Process Bionics