A Proposal to Model Knowledge in
Knowledge-Intensive Business Processes
Norbert Gronau, Christoph Thim, André Ullrich, Gergana Vladova and Edzard Weber
Business Informatics, esp. Processes and Systems, Potsdam University, August Bebel Str 89, Potsdam, Germany
Keywords: Knowledge, Business Process, Ability to Articulate, Generality, Information.
Abstract: This paper addresses the topic of modeling tacit knowledge across business processes. Some approaches exist
to cover that issue but none is really satisfying. Therefore a new approach is proposed, which is based on
more than ten years of experience with the Knowledge Modeling and Description Language (KMDL). The
new approach suggests to differentiate knowledge in professional insight, experience and context and to
describe the degree of ability to articulate and generality.
1 INTRODUCTION
This paper addresses the topic of modeling tacit
knowledge across business processes. Some
approaches exist to cover that issue but none is really
satisfying. Therefore a new approach is proposed,
which is based on more than ten years of experience
and also overcomes the deficits of existing
approaches.
One of the often used definitions of tacit
knowledge is based on Davenport‘s set of criteria that
consists of information, professional insight, values,
experience and context (Davenport and Prusak, 1998;
Gronau, 2012). Conventional approaches for the
differentiation of knowledge types like Polanyi
(1966) or for the differentiation of the handling of
knowledge like the SECI model (Nonaka and
Takeuchi, 1995) see tacit knowledge only bound to
humans. That might be too narrow in the light of new
cyber-physical systems as self-organizing and
decision-capable technical entities (Lee et.al, 2014;
Gronau, 2015). In the future at least some of the
competencies to make decisions will lie with
technical actors.
Digitalization, virtualization and the Internet-of-
things force great changes in the roles of the
employees and the technical actors. Machines and
factory units collect data from their environment with
the help of sensors, process these data and act in the
environment using mechanical actuators. Data will be
sent to information systems, which receive, process
and forward them. This is an analogy to the human
information processing. Processing includes the use
of information following predefined rules and a
predefined space of alternative solution paths
(Inference), and the creative development of facts and
solutions additionally to predefined structures with a
not predetermined result (intelligence, cf. Turing,
1950).
Knowledge as a „goal-oriented netting of
information“ (Rehauser and Krcmar, 1996) allows
that actors to act and to decide. It helps to prepare
decisions and is an important component to generate
competencies. The netted information contains data
with semantics and data with a certain syntax. Human
as well as technical actors are able to proceed signs,
data information and knowledge with existing
technology. Therefore it might be useful to see also
the technical entities as potential bearers of
knowledge. While value creating processes become
more and more interwoven with cyber-physical
systems, some of the concepts developed for person-
bound knowledge also can be used for a machine‘s
knowledge. Especially the aspects of professional
insight in a specific domain and the experience are
candidates for a transfer from man to machine.
Experience for instance can a machine gain and
process by using a case-based-reasoning system.
Another problem occurs when the usage of
knowledge in teams is investigated. This kind of
knowledge cannot be characterized with the criteria
given by Davenport, because most of these criteria are
only suitable for one human.
What makes the difference between subjective
knowledge and explicable and more objective
98
Gronau N., Thim C., Ullrich A., Vladova G. and Weber E.
A Proposal to Model Knowledge in Knowledge-Intensive Business Processes.
DOI: 10.5220/0006222600980103
In Proceedings of the Sixth International Symposium on Business Modeling and Software Design (BMSD 2016), pages 98-103
ISBN: 978-989-758-190-8
Copyright
c
 2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

knowledge? It is the context, in that the information
is announced and it is the value that may be very
different for two humans looking at exactly the same
piece of information. Therefore the authors of this
contribution see information as an extreme
occurrence of knowledge. During the process of
explication the context was reduced and the value
propositions were omitted.
One of the modeling techniques that is able to
represent knowledge bound to persons is the
Knowledge Modeling and Description Language
(KMDL®). Its development started more than ten
years from now. In that time a lot of experience was
gained, especially in the areas of software
development, product development, innovation
processes, quality management and other areas
(Gronau, 2012). Based on these experiences the
authors suggest to differentiate knowledge following
the criteria of professional insight, experience and
context and to look at generality and ability to
articulate for each of these criteria. The following
sections describe this proposal in more detail.
2 THE TERM KNOWLEDGE
Stemming from the complexity of the term
knowledge the necessity occurs to differentiate in
knowledge types and knowledge dimension. The
supposedly most often used differentiation
discriminates between tacit and explicit knowledge.
The tacit dimension was first described by Polanyi
and addresses parts of the personal knowledge, which
are neither to be scribed nor to be articulated.
„Although the experts (...) can indicate their
clues and formulate their maxims, they know
many more things t ha n th e y c a n t e ll, knowing
them only in practice, as instrumental
particulars, and not explicitly, as objects.“
(Polanyi 1958, S. 88)
Tacit knowledge is „personal, context specific and
very difficult to communicate“ (Nonaka and
Takeuchi 1995, p. 72). Contrarily explicit knowledge
can be distributed in a formal and systematic
language. Tacit knowledge can be seen as a synonym
of embodied and procedural knowledge (Meyer and
Sugiyama, 2007, p. 26).
Davenport and Prusak (1998, p. 5) deliver a so-
called pragmatic definition of knowledge:
„Knowledge is a fluid mix of framed
experience, values, contextual information,
and expert insight that provides a framework
for evaluating and incorporating new
experiences and information. It originates and
is applied in the minds of knowers. “.
Knowledge is seen as very difficult to articulate
and also person-bound. It is based on information but
cannot be equaled with it. To make the term
knowledge more comprehensible, Davenport and
Prusak (1998) refer to six key components:
experiences, ground truth, complexity, judgment,
rules of thumb and intuition, values and beliefs.
Explicit and tacit (some authors use the wrong
term of „implicit“) knowledge are defined by pointing
out the difference in processing these two knowledge
types. Explicit knowledge can be transferred by
communication, by numbers, pictures or language. It
can be processed, altered and learned together
(Willke 2001; Franken and Franken 2011, p. 33).
Lam (2000) has given a description of knowledge
that refers not only to qualities but also to the
organizational context: The encoded knowledge has
an existence independent of persons and can be stored
in handbooks, data bases, rules of conduct etc. and
can be seen as organizational explicit knowledge (see
also Blackler, 1995). The embedded knowledge to the
contrary cannot be transferred objectively but is
socially constructed, captured in organizational
cultures, language systems etc and used and shared by
the members of the organization. Different types of
knowledge are differentiated in the realm of
organizational knowledge:
 encultured knowledge, which is shared by the
members of the organization and transferred by
socialization (Sackmann, 1991; Kogut and
Zander, 1992)
 event knowledge that is concerned to events in the
lifetime of the organization (Vlaar et al, 2007)
 procedural knowledge about processes and
connections (Fischer, 2008).
 embodied knowledge describes the dimension of
individual tacit knowledge. It is bound to persons
and can only be created by experience (Polanyi,
1966; Blackler, 1995; Nonaka and Takeuchi,
1995).
Franken and Franken (2011, p. 30) say that
knowledge is something immaterial, difficult to
describe, but with great influence on human acting. It
has to be distinguished between the real world on the
one hand and the immaterial world of knowledge on
the other hand, which exist in the human brain as a
result of experiences and learning, leading to mental
patterns. In this way knowledge is developed as an
A Proposal to Model Knowledge in Knowledge-Intensive Business Processes
99

individual construction from the interaction with the
real world (Franken and Franken, 2011, p. 31).
Rehauser and Krcmar (1996) denote knowledge as an
individually modeled reality, which is generated by
the bearer of knowledge under the influence of her
own perspective. Knowledge allows to act and the
artifacts created during the action cause a change in
the real world, induced by the individual person.
Summarizing it can be reasoned that the term
“tacit knowledge” encompasses a broad area of
different characteristics. This makes its transfer into a
model, which is necessary to get a grip on knowledge
processes and knowledge conversions, very difficult.
In the following sections the deduction of such a
model in the context of KMDL is presented.
3 DEDUCTION OF A CONCEPT
FOR THE MODELING OF
KNOWLEDGE
Staring with the different characteristics of the term
tacit knowledge a classification is necessary as a first
step. First knowledge can be classified following the
definition of Davenport and Pruzak (1998, p. 5).
Following them knowledge consists of experience,
values, context information and professional insight.
Experiences stem from a practical engagement
with a certain topic. Professional insight is the
intellectual penetration of an area of content. Values
are generated by socialization procedures and are
shared conjointly. They are deeply embedded into
one‘s personality. On the other side context
information is the picture of an observation. This
observation can relate to an object, a person, a topic
of the environment or a self-observation.
Existing approaches that recognize these
differentiations and the relations between the
components are rare (Hinkelmann et al, 2002; Heisig,
2000; Allweyer, 1998; Gronau and Froeming, 2006).
Following the guidelines of proper modeling (Becker
et al, 1998) the possibility and usefulness of every
component has to be judged. The six requirements are
relevancy, correctness, economic feasibility,
clearness, comparability and systematic construction.
The question of relevancy was solved by selecting
criteria and justify their selection above. The other
requirements are combined to judgment criteria.
The requirements correctness and economic
feasibility are merged into the criterion
ascertainability. Ascertainability states whether
components of knowledge can be grasped objectively
true and whether this is possible with reasonable
effort. A modeling of a component of knowledge is
only possible if this component can be captured by an
observer or by self-observation.
The requirements clearness and comparability are
combined into the criterion intersubjective
comparability. The intersubjective comparability also
is important to be able to compare certain model
statements and to be able to model the transfer of
knowledge.
These components of knowledge can be captured
and compared in different degrees of easiness, as seen
in Table 1.
Table 1: Judgment of easiness of modeling.
Component Ascertainability
Intersubjective
comparability
Professional insight ++ ++
Experience ++ ++
Values o --
Context + +
++ very good, + good, o no statement possible, - bad, --very bad
Professional insight, for instance in the shape of
formal education, can be captured by certificates or
the documentation of training periods. These are also
comparable very good, by certificate degrees, age of
knowledge etc.
Experience can be captured objectively by
documenting core areas of action or by self-judgment.
Although distortions are possible, typically the results
are mostly correct. Also an intersubjective
comparability is given, when durations, frequency or
intensity of actions are compared between different
knowledge bearers.
Values are very difficult to capture due to their
often un-reflected anchoring in the human
consciousness and their very subjective character.
Also an intersubjective comparison between values is
not possible, because it is very difficult to create a
hierarchy of values or to compare the value systems
of two humans. A pure description of equal or
different values is not suitable for the modeling
purpose. Another argument is that the dissemination
of values in an organization occurs over time and is
of long duration, therefore not usable in the context
of process-oriented knowledge management.
Properties of values that are relevant for decisions can
be modeled in the context component.
The context component can be captured in a
sufficient manner when the usage environment is
Sixth International Symposium on Business Modeling and Software Design
100

described or observed. Although the context can be
compared inter-subjectively, different interpretations
or perspectives can occur.
Following those thoughts, values cannot be
modeled sufficiently. The remaining components to
model knowledge objects are therefore professional
insight, experience and context.
Knowledge seldom can be assigned only o one
single component. The judgment of the context
typically uses experience. Capturing of professional
insight is done within a context and the collection of
experience only works when professional insight is
available. Therefore these components have to be
inspected together depicting the knowledge of person,
an item or a status.
Beside the differentiation of knowledge
components to be able to model the use and the
transfer of knowledge more information is necessary.
For a more detailed description the knowledge
dimensions of Spinner (2002) can be used. He
differentiates the shape, expression, content and
validity dimensions.
Table 2: Judgment of modeling of knowledge dimensions.
Component Ascertainability
Intersubjective
comparability
Shape ++ ++
Expression + +
Content + --
Validity - +
++ very good, + good, o no statement possible, - bad, --very bad
The shape dimension indicates the generality of
knowledge on a scale between particular and general.
The expression dimension depicts the degree of
articulation and has the polar characteristics tacit and
explicit. The content dimension indicates how much
information lies in the knowledge, between nearly
and highly informative. The validity dimension
shows how much the knowledge is backed by facts or
scientific results. This dimension has the polar
characteristics of hypothetical and apodictic.
Again the dimensions can be checked with their
degree of ascertainability and intersubjective
comparability to judge the transfer into knowledge
modeling (Table 2).
Following Table 2 we can see that especially the
dimensions of shape and expression are suitable to
integrate into modeling. The content dimension
cannot be compared inter-subjectively, due to
different prevalent knowledge and different interest
in the subject. The content dimension is therefore
different between two persons and during different
points in time. Additionally no judgment of the value
propositions of the bearer of knowledge is intended,
especially because it is very difficult to measure a
value proposition. Nevertheless the authors ant to
state that the attached value remains an important part
of the description of person-bound knowledge.
Further on the validity dimension is difficult to
capture on an individual level. Whether some element
of knowledge is hypothetical or rock solid cannot be
determined in most cases.
The concentration of the two remaining
dimensions allow a more detailed description of
knowledge. Both dimensions can be applied on the
components so that a 2x3 matrix is constructed (Table
3).
Table 3: Characteristics of a knowledge object.
Professional
insight
Experience Context
Ascertainability [0,1] [0,1] [0,1]
Generality [0,1] [0,1] [0,1]
Professional insight, experience and context are
judged referring to generality and ascertainability by
the bearer of the knowledge with values from 0 to 1.0
means, there is no expression of this characteristic
while 1 means there is a maximal expression of this
characteristic. For the dimensions it means as
follows:
Ascertainability:
0 - not articulable, real tacit knowledge
1 - completely articulable
Generality:
0 - particular, only useful in a single instance
1 - commonly useful
Instead of the suggested numbered scales also other
scales are possible, so for instance pure yes-no-
depictions or judgments like low - medium - high.
Using these six characteristics, a very detailed
differentiation of a knowledge object can be
processed. Therefore it is suggested to use this new
knowledge object while modeling with KMDL
(Figure 1).
Figure 1: Knowledge object in KMDL.
A Proposal to Model Knowledge in Knowledge-Intensive Business Processes
101

Figure 2: Modeling of internalization of knowledge.
After introducing the multi-dimensional of
knowledge objects a differentiation between
knowledge and information objects can be omitted.
Therefore in the activity view of MDL only
knowledge objects are shown and the information
objects move to the process view to assure
comparability to other BPM modeling approaches.
An additional advantage lies in the better ability
to interpret the conversion of knowledge. This is
explained using two examples:
Example 1: ERP usage in chemical industry
An expert of ERP systems in the chemical industry
can articulate her knowledge with a degree of
ascertainability of 0.85. Under some circumstances
she will get her knowledge about that topic from
books and journal essays, but not only from practical
experience. Therefore she has a great ability to
articulate but a quite limited experience.
Example 2: Vegan food
Now the same expert from example 1 shall speak
about vegan food. Due to missing personal
experiences but because the ascertainability of the
expert her knowledge can be assessed, Low values for
generality mean that her knowledge is not very useful
for others, although she is able to articulate it quite
good.
Beside the better representation of the knowledge
of certain actors in the process also the knowledge
conversions externalization, internalization
socialization and combination ca be represented
better. The modeler has to decide about his point of
observation and about the purpose of the modeling
beforehand. By comparing the scale expressions of
the bearer of the knowledge object before and after
the conversion also an increase of knowledge can be
measured - clearly a real advantage against other
modeling approaches!
In Figure 2 an internalization is depicted using the
newly developed knowledge object. Not the transfer
of knowledge from the printed dissertation to the
knowledge of he bearer is of interest here but the
increase of the bearer‘s knowledge about the topic
before and after reading the dissertation. This can be
seen that in four of the six characteristics of a
knowledge object an increase took place and only two
characteristics remain unchanged.
4 CONCLUSIONS
Modeling the occurrence of knowledge is the decisive
key to be able to recognize potentials for the
improvement of knowledge-intensive business
processes. For this purpose a differentiation of
knowledge is very important. This paper proposed a
framework to capture only these dimensions of
knowledge which can truly being captured during
modeling.
ACKNOWLEDGEMENTS
The authors would like to thank Marcus Grum for his
many valuable suggestions.
REFERENCES
Allweyer T, 1998. Knowledge Management with ARIS
Models. In: Scheer A-W: ARIS - From Business pro-
cess to the application system (in German), Heidelberg
Sixth International Symposium on Business Modeling and Software Design
102

Becker J, Rosemann M; Schuette R, 1995. Guidelines of
proper modeling (in German). In:
Wirtschaftsinformatik (37) 5, 435-445.
Blackler F, 1995. Knowledge, Knowledge Work and Or-
ganizations: An Overview and interpretation; in:
Organ-ization Studies, I6(6),1021-1046.
Davenport T, Prusak L, 1998. Working knowledge: How
organizations manage what they know, Harvard Busi-
ness School Press, Boston.
Fischer M, 2008: Organizational learning in Corporations
with extreme performance pressure (in German). In:
Pawlowsky P; Mistele P (Eds.): High performance
management (in German), Wiesbaden: Gabler,lll-131
Franken R, Franken S, 2011. Integrated knowledge and
innovation management (in German). Wiesbaden: Ga-
bler.
Gronau N, 2015. The influence of Cyber-Physical Systems
on the creation of production systems (in German). In-
dustrie Management 3, 16-20
Gronau N, Heinze P, Weber E, 2012: Conversion- and
Transformation-Oriented View on Knowledge Flows.
In: Gronau N (ed): Modeling and Analyzing knowledge
intensive business processes with KMDL - Compre-
hensive insights into theory and practice, Berlin, 249–
268.
Gronau N, Fröming J, 2006: A semi-formal modeling lan-
guage to describe knowledge conversions (in German).
In: Wirtschaftsinformatik (48) 349-360.
Heisig,P, 2002. GPO-WM - Method and tool for process
oriented knowledge management (in German). In:
Abecker A, Hinkelmann K, Maus H, Mueller HJ(eds.):
Process-oriented Knowledge management (in German),
Berlin Heidelberg, New York, 47-64.
Hinkelmann K, Karagiannis D, Telesko R, 2002. PRO-
MOTE - Method and tool for process oriented
knowledge management (in German). Abecker A, Hin-
kelmann K, Maus H, Mueller HJ (eds.): Process-
oriented Knowledge management (in German), Berlin
Heidelberg, New York,
Kogut B, Zander U, 1992. Knowledge of the Firm: Combi-
native Capabilities and the Replication of Technology;
in: Organization Science, Vol. 3, No. 5 (August), 383-
397
Lam A, 2000. Tacit Knowledge, Organizational Learning,
and Societal Institutions: An Integrated Framework.
Organization Studies, vol. 21 (3), 487-513.
Lee J, Bagheri B, Kao H-A, 2014. A Cyber Physical Sys-
tems Architecture for Industry 4.0-based
Manufacturing Systems. Manuf Lett 3, 18–23
Meyer B, Sugiyama K, 2007. The concept of knowledge in
KM: a dimensional model", Journal of Knowledge
Management (11) 1, 17 - 35
Nonaka I, Takeuchi H, 1995. The knowledge creating com-
pany: how Japanese companies create the dynamics of
innovation. New York: Oxford University Press, 284
Polanyi M, 1966. The tacit dimension. Garden City, Dou-
bleday
Polanyi, M, 1958. Personal Knowledge: Towards a Post-
Critical Philosophy. University of Chicago Press, Chi-
cago, IL.
Rehauser J, Krcmar H, 1996. Knowledge Management in
the corporation. In: Schreyoegg G, Conrad P (eds.)
Management research 6: Knowledge Management.
Ber-lin, 1-40
Sackmann S, 1991: Uncovering Culture in Organizations;
in: Journal of Applied Behavioral Science, 27, 295-317.
Spinner HF, 1994. The knowledge order. A guiding concept
for the third order of the information age (in German)
Opladen
Turing AM, 1950. Computing machinery and intelligence.
Mind, 433-460.
Vlaar P, Van Den Bosch F, Volberda H, 2007. Towards a
Dialectic Perspective on Formalization in
Interorganiza-tional Relationships: How Alliance
Managers Capitalize on the Duality Inherent in
Contracts, Rules and Proce-dures; in: Organization
Studies, Vol. 28, 437 - 466
Willke H. 2001. Systemic Knowledge Management (in
German). Stuttgart. 2. edition.
A Proposal to Model Knowledge in Knowledge-Intensive Business Processes
103