Sunday, November 8, 2009

Process Mining, Knowledge Creation, Ontologies and Frameworks

Can these capabilities be used for a better understanding of ERPs?

This article discusses how these “capabilities” can be combined to facilitate the use of ERPs.


Process mining allows the reverse engineering of event logs into workflows that describe how a process behaves. Traditional process mining techniques are based on syntax. Typically process mining requires; a unique process id, event names and types, date and time and user names.

Most ERPs are not suited to this kind of analysis. They are either without any accessible process id or a reliable audit trail. However, applications running on the IBM i (previously AS400 or iSeries) have a journaling capability that can provide an answer to this problem.

Using semantic knowledge allows us to extend the process mining capability and explore other disciplines. ERPs are rich in knowledge and meaning. There are structures that are known to IT users but often not well understood by the business. One of the issues that business users have with ERPs is their complexity. An objective must be to introduce greater clarity and simplicity to the ERP.

Capabilities such as the use of ontologies and standard frameworks (APQC PCF) create the structure that allows semantic process mining and the creation of new knowledge.

These capabilities allow an organisation to ask new questions;
How is SAP Purchasing different to other ERPs?
What capabilities does SAP have compared to other ERPs?

Knowledge Creation

One of the difficulties in creating new knowledge is the conversion of tacit knowledge to explicit knowledge. Tacit knowledge is about know-how. It involves mental models. The sort of knowledge that organisation find hard to document and share within an organisation. This is true of application systems such as an ERP.

For example, every ERP has a Purchasing module that allows the organisation to purchase materials and manage relationships with vendors. There may be a set of applications that form a process called “Process Purchase Orders”. However this knowledge may not be readily available to business users. One way to resolve this is by building ontologies. This will allow the creation of explicit knowledge that can be shared and understood throughout the organisation.

AQPC Process Classification Framework

“Created in collaboration with expert and practitioner panels in 1992 and updated annually, APQC's Process Classification FrameworkSM (PCF) is a globally recognized business process model that defines activities and processes across 12 enterprise-level operating and management categories.” (APQC)

The PCF is also a good benchmark to make sure you haven’t missed anything when reviewing an ERP. Any process that is discovered should exist by a similar name in the PCF. See reference for PCF details.


The basis of semantic business process mining is the ontology. Ontologies can represent applications, organisation or business objects.

There are two main benefits in using this kind of ontology. Firstly, workflows are created the mid level descriptors (Change Purchase Order etc) can be used rather than the codes. Secondly, like processes are grouped in one place so that business users can see the applications capabilities.


These capabilities are useful on their own but are much more powerful when used together. If the goal is knowledge creation, process mining provides new information that is organised through frameworks and ontologies. If the goal is business process management, process mining provides discovery, the ontologies provide clarity and structure and the frameworks provide a checklist and a standard.

The Knowledge Creating Company; Nonaka; Takeuchi, OUP 1995

1 comment:

Sweety_Fatty143 said...

How can Business Process Management optimize business outcomes and performance?

bpm business process