Scheer
Scheer
Scheer Wiki

Menu

Wiki

23.09.2019

Artificial intelligence as a driving force of smart business process management

Artificial intelligence (AI) has already found its way into many areas of daily life. Following frequent application, AI is often no longer seen as such but rather as functioning little helpers in daily life. Life is now practically inconceivable without navigation systems and language assistants as well as the automatic classification of vacation photos on the smartphone, for example.

Apart from such little daily helpers however, artificial intelligence can also provide great added value within companies, specifically within the execution of business processes. In the course of daily growing requirements and increasing cost pressure, companies need to establish new concepts for increasing efficiency accompanied by the automation of routine activities. But how can a company efficiently identify, exploit, and then, in a purposive manner, integrate such potential in the operationally important processes of added value?

Generally called for are the latest technologies for analyzing vulnerabilities within a company, software robots for automating routine tasks, and automated decision-making based on artificial intelligence for dedicated use cases. Today, these aspects are part of every strategic corporate orientation pursuing growth and resource optimization.

As soon as it works, nobody calls it AI anymore
– John McCarthy

Intelligent robotic process automation

In combination with AI technologies, robotic process automation will in the future contribute to the fully automatic completion of simple, frequently recurring tasks. These intelligent software robots already have an enormous influence on many existing processes in the current working world. Simple processes, as in clerical processing, which can be found in any company in the commercial sector or in public administration, will in the future be automated through AI-aided software robots, so as to be executable around the clock without downtimes.

Given the large number of cases in the standard processes, enormous resources will be freed for more sensible utilization in special clerical processing, for example. Intelligent software robots combinable in networks consequently contribute to process automation both effectively and at low cost.

Enormous gains in efficiency can then be achieved for occupational groups whose work consists in classical clerical processing on the one hand and supporting customers on the other, as in the banking sector. Simple process steps, such as the release of entries in critical account balances, can already be completed automatically – decision-making is enabled by so-called "weak" AI.

DIGRESSION: Week AI vs. strong AI

Contrary to expectations, artificial intelligence today can solve only small dedicated problems in an automated way. As a simple example, artificial intelligence designed for voice recognition can tackle only this particular task and cannot be used for predicting possible subsequent process steps or future sales figures. These trained AI services, functioning well in themselves, thus belong to the week AI category. In the particular context of business, such decision-making aids can already lead to enormous improvements efficiency today. So-called "strong" AI, which can react to different issues individually and in a causal way, is still a subject of current research.

The consequently freed capacities become available in the relatively more intensive area of customer care to promote the essential goals of the company. At present there are numerous examples of clerical processing being aided by software robots. But it quite often happens that AI techniques are necessary for achieving the desired level of automation. The further use of constantly improving AI methods will increasingly contribute to automating additional and more complex matters as well.

In the future, digitalization strategies will be determined by new software concepts for corporate management. They include robotic process automation and process mining, combined with artificial intelligence.
– Prof. Dr. Dr. h.c. August-Wilhelm Scheer

Intelligent process mining

Process mining constitutes a special discipline within data science. Specifically, algorithms are applied to so-called log files, which reflect the sequence of processes in a workflow or ERP system. Actual processes can be derived from the systems in an automated manner and checked for efficiency, among other things. In this way, actual processes from a system can be visualized at the press of a button, to provide a starting point for process automation projects (as in robotic process automation) or system transformations (when switching from SAP R3 to S/4-HANA, for example). Three different types of analyses can be distinguished in classical process mining (see Dadashnia 2017):

  • Discovery
    Discovery is the procedure that uses an event log file as a basis to generate a process model via dedicated algorithms. Process discovery is a technique commonly used in process mining, being an effective procedure for surveying the business activities and processes actually taking place in the company.
  • Conformance Checking
    Conformance checking compares a pre-existing process model (for example, a target model) with a model captured through process mining on the basis of event logs (actual model). The method detects deviations within defined and actually running processes.
  • Enhancement
    The basic mode of operation in this approach lies in enriching existing modeled processes with relevant information from analyses of the event logs, and thereby improving the model.

Current developments and the use of AI algorithms will allow expanding previous process-mining analyses in the future. For example, not only can we retrospectively view currently running business processes after their completion, we can make appropriate recommendations for action directly during the execution of the processes. These recommendations can help towards detecting and correcting production errors even during production, for example. Such an approach is also known as in-instance process improvement.

The continuously growing volume of data and increasing computing power, accompanied by better and better AI algorithms, combine to improve past methods and concepts in business processes management, though partly with disruptions. Given the high demand and the demonstrated applications, intelligent business process management offers strong use potential and will increasingly enter into business processes in all sectors.
– Sharam Dadashnia