A FRAMEWORK FOR SUPPORTING
KNOWLEDGE WORK PROCESSES
Weidong Pan, Igor Hawryszkiewycz and Dongbei Xue
Faculty of Information Technology, University of Technology, Sydney, PO Box 123 Broadway, NSW 2007, Australia
Keywords: Knowledge work processes, supportive services, workspaces, software agents, description of knowledge
work processes, best practices, cost-effectiveness.
Abstract: Improving knowledge work processes has become increasingly important for modern enterprises to
maintain a competitive status in nowadays information society. This paper will propose a way to improve
knowledge work processes through supportive services. A framework for supporting knowledge work
processes will be presented where the best practices of knowledge work processes, developed by process
organizers or derived from some successful applications, are described and stored in a database, and
according to the description, software agents dynamically organize supportive services to guide process
participants to advance process steps towards the efficient completion of a process. The paper will provide
an overview of the method and explore the development of the main components in the framework.
1 INTRODUCTION
With the shift from an industrial economy to a
knowledge based economy, improving knowledge
work processes has become increasingly important
in modern enterprises for providing competitive
advantage and sustaining their success (Drucker,
1999). Knowledge work processes are knowledge-
intensive business processes, where the dominant
process activities involve abstract reasoning,
problem-solving and decision-making. Examples
include market research and development, strategic
planning, new product development, etc. To achieve
a quality outcome, most such processes require their
participants to have sufficient information, specialist
knowledge and experience and also require them to
mutually collaborate and cooperate as well.
Currently participants have to carry out considerable
amount of manual work to search, assemble and
organize various required information and develop
contacts to achieve an outcome.
Searching ways to organize knowledge work
processes with minimal manual work and lowest
cost has attracted many research efforts. The
problem is that most of them have been focused only
on information management (Röll, 2004). Very little
has been done to improve the entire process. There
is no integration of technologies to support the
different dimensions of a knowledge work process.
The ways for how to facilitate the completion of a
knowledge work process with higher efficiency and
lower cost are still lacking to date.
In this paper we will propose a way to improve
knowledge work processes through providing
supportive services for participating actors. A
framework for supporting knowledge work
processes will be presented. A work process
specification language (WPSL) will be developed to
describe best practices of knowledge work processes,
developed by process organizers or derived from
some successful applications. A multi-agent
architecture will be developed to guide participants
to advance process steps. The software agents in the
architecture will, according to the knowledge
captured from the description of the best practices,
organize and provide services for participants to
facilitate process completion with higher efficiency
and lower cost. The paper is organized as follows.
Section 2 provides a general description of a
knowledge work process. Section 3 presents our
research strategies to improve knowledge work
processes, where a framework for supporting
knowledge work processes is proposed. Section 4
deals with the implementation of workspaces.
Section 5 explores the development of the database
storing best practices of knowledge work processes,
252
Pan W., Hawryszkiewycz I. and Xue D. (2007).
A FRAMEWORK FOR SUPPORTING KNOWLEDGE WORK PROCESSES.
In Proceedings of the Ninth International Conference on Enterprise Information Systems - DISI, pages 252-257
DOI: 10.5220/0002349702520257
Copyright
c
 SciTePress

described by using the WPSL. Section 6 addresses
the development of software agents realizing the
supportive services. Finally, section 7 summarizes
the paper with an outlook on the further work.
2 GENERAL DESCRIPTION OF A
KNOWLEDGE WORK
PROCESS
Knowledge work processes are a category of
business processes. A business process involves
many different dimensions and hence, there are
various possible ways to classify business processes.
Figure 1 shows a classification of business processes
by the process complexity dimension and the
knowledge intensity dimension (Mertins et al.,
2001). It maps business processes into four classes.
Examples for each class are respectively put into its
corresponding quadrant. Our research is aimed to
support the business processes that are mapped into
the right part of the figure and in particular the ones
in the third quadrant.
Knowledge work processes not only include
structured process routines but also include various
knowledge-intensive activities, such as reasoning,
problem-solving, decision-making, etc. In doing
these knowledge-intensive activities, knowledge
workers access data, use knowledge, employ mental
models, and apply significant concentration and
attention (Davis, 2002). Obviously knowledge work
has distinctive characteristics, which differ from
administrative, operational work, or other structured
work. The typical ones can be summarized as
follows: 1) it requires participants to possess a high
level of skill and expertise; 2) it often involves
distributed work and requires group efforts to
accomplish; 3) there are no predicted task structures
and sequences and their evolution may be influenced
by the outcomes of the earlier stage; and 4) there are
no standard solutions and they may vary somewhat
based on the actor doing the work. To characterize
these characteristics in combination, emergent might
be a good label (Markus et al., 2002).
Figure 2 provides a general description for a
knowledge work process (Hawryszkiewycz, 2006).
It can describe many concrete processes in modern
enterprises, such as improving the quality of a
product in a company, for example. In such a
process, a situation is identified at first, for example
by receiving feedback from customers on the
product quality. Then the situation is assessed and
alternate courses of action are then determined based
on the assessment and other knowledge relevant to
the product and the company. A suitable course of
action is chosen from them. The relevant execution
plan is designed and then that course of action is
executed. These process steps are performed by
different actors who may work at different places.
As seen from Figure 2, an actor, to take his/her
role, not only needs knowledge and experience for
knowledge intensit
y
Strong Weak
process complexity
low high
①
②
④
③
Figure 1: A classification of business processes.
order configuration
new product sign-of
f
editing
order fulfillment
yield analysis
advertising
product innovation
new product development
information management
new business development
marketing r & d
capital projects
complaint handling
customer service
p
lanning auditing
leasing lending
Figure 2: A general description of a knowledge work process.
Selecting a course
of action
Execution
p
lannin
g
Task
leade
r
Defining alternate
courses of action
Local
manager
Organizational
director
y
Assessor
External
contact
Identifying
a situation
Situation
assessmen
t
Situation
description
Unit
manage
r
Execution
Courses of
action
Execution
management
Coordinator
A FRAMEWORK FOR SUPPORTING KNOWLEDGE WORK PROCESSES
253

his/her own responsibility but also requires
information pertaining to other actor’s work and
cooperation with other actors for a common goal.
This paper suggests software agents be integrated
into the work environment to help them get such
information and guide them to get through process
steps. In the following sections, the description of a
knowledge work process will be used as a basis for
our research to improve knowledge work processes.
3 AN AGENT BASED WAY TO
IMPROVE KNOWLEDGE
WORK PROCESSES
The research presented in this paper aims to develop
supportive services to improve knowledge work
processes toward best practices based on the cost-
effectiveness principles. It will develop a structure
like that shown in Figure 3 to assist participants
working across distances to conduct process
activities and facilitate process completion with
higher efficiency and lower cost. More specifically,
• It will develop electronic workspaces for
participants to conduct process activities,
exchange knowledge and information. A wide
range of supportive tools, including knowledge
mining tools, information visualization tools,
communication and collaboration tools, will be
seamlessly integrated with the workspaces to help
the participants capture and annotate information,
make sense of information, reflect, get feedback,
share, discuss and cooperate with others across
distances (Röll, 2004). The supportive services for
facilitating process completion will be provided
for the participants also through the workspaces.
• It will develop a database, process guide database,
to store various “best” approaches to performing a
knowledge work process. Following them,
participants may capture best practices in
accomplishing the process. To develop such a
database, it is required to analyze and examine
knowledge work processes to structure process
activities and build a metamodel of “best”
progression by using related theories. Moreover, it
is required to describe the metamodel in a formal
way. Using the knowledge captured from the
description stored in the database, agents will
guide different actors in Figure 2 to advance
process steps towards an efficient completion of
the knowledge work process.
• It will develop a multi-agent architecture to
actively organize and provide services for
participants in the given situations. The agents will
work at the background, observe and monitor the
events that happen in the workspaces, and based
on the actual progression at execution-time, enact
process activities, manage them, and provide
supportive services to facilitate their efficient
completion, by using the knowledge captured
from the process guide database. All the services
will be customized to individual participants based
on their particular needs.
In short, our framework mainly includes three
components: 1) the electronic workspaces for
conducting process activities, 2) the process guide
database storing best practices, and 3) the multi-
agent architecture realizing supportive services
according to the best practices stored in the database.
In the following sections, we will respectively
explore the development of these components.
Figure 3: Overall architectural framework for supporting knowledge work processes.
Supportive
services
Monitoring
Electronic
workspaces
Multi-agent
architecture
Process guide
database
……
Knowledge mining tools
Visualization tools
Collaboration tools
Supportive tools
Process participants
who may work at
different locations
External
contact
Local
manager
Assessor
Coordinator
Task leader
Unit manager
……
Use the
knowledge
in the
database
Use a tool
The knowledge work process support system
Integrated with
Interactions
Interactions
ICEIS 2007 - International Conference on Enterprise Information Systems
254

4 DEVELOPMENT OF THE
WORKSPACES
Since knowledge and information for a knowledge
work process is typically distributed over different
locations, participants require knowledge sharing to
achieve an outcome (Markus et al., 2002). It is thus
necessary to provide a space over the Internet for
them to carry out relevant process activities and
implement required collaboration and knowledge
sharing. Electronic workspaces are such a shared
mental space implemented by advanced ICT, where
participants can not only exchange views and ideas
but also share feelings, emotions, experiences, and
other mental models. On the workspaces, they can
work individually, and if necessary, they can work
collaboratively with others towards a common goal.
They are given varied flexible and convenient means
to support various process activities.
The electronic workspace shown in Figure 3 can
be implemented by our LiveNet system
(Hawryszkiewycz, 2005), developed based on the
requirements for supporting collaborative process
processing. It supports both synchronous and
asynchronous collaboration of distributed groups of
participants through utilizing workspaces as the
commonly shared place. By LiveNet workspaces,
participating actors, no matter where they are, are
able to share information with one another in their
own workstation, for example, they can use a same
document or discuss a same issue. They can
exchange messages one another by text or graphic
images. Also, it provides a diversity of technological
facilities and means for facilitating process
progression through workspaces.
In LiveNet, any number of roles can be defined
for a workspace, and users can be added to a
workspace as participants taking roles. One user can
have different roles in different workspaces, but can
only occupy one role in each workspace.
LiveNet provides the flexibility to set up
workspaces and supports their dynamic evolution as
the process emerges. An example is shown in Figure
4, which illustrates workspace evolution in a process
for improving the quality of a product in a company.
As depicted in the left screenshot of the figure, the
workspace named as improving product quality
contains two projects (sub workspaces) and three
artifacts, five roles have been defined for the
workspace, and four users have participated in
activities within the workspace. The right screenshot
shows the workspace after the process progresses. It
can be seen from the screenshot that four artifacts
have been added to the workspace.
Figure 4: Progression of workspaces in a knowledge work process.
A new plan for
keeping track
of progress
A new text
file for
sharing
A new
calendar for
recording
availability
A new folder
for sharing
Workspace name
Tools bar for
workspace evolution
A FRAMEWORK FOR SUPPORTING KNOWLEDGE WORK PROCESSES
255

5 DEVELOPMENT OF THE
PROCESS GUIDE DATABASE
To guide participants to advance process steps in
knowledge work processes towards an efficient
completion by using the structure shown in Figure 3,
a principal challenge is to provide knowledge for
software agents so that they can support all the
different activities shown in Figure 2. It is thus
required to explore ways to evolving process steps
towards the attainment of the greatest process
productivity and lowest cost, and develop a formal
way to describe these ways. A work process
specification language (WPSL) is thus necessary to
describe knowledge work processes.
Building upon our earlier work on facilitating
learning processes (Pan & Hawryszkiewycz, 2006a)
and a hypothesis that a knowledge work process can
be described and facilitated in a similar way as a
learning process, we have designed a WPSL for
specifying knowledge work processes (Pan &
Hawryszkiewycz, 2006b). Consequently the best
practices of knowledge work processes, developed
by process organizers or derived from some
successful applications, can be described using the
WPSL in a formal way and stored in a process guide
database. This section will investigate into the
development of the database.
5.1 Requirements for the Database
The purpose for developing the process guide
database is to store ways to evolving process steps
in a knowledge work process so that agents can
capture knowledge for supporting process activities
and providing guidance for individual participants to
facilitate the completion of process activities in a
robust way. As shown in Figure 5, the database
provides links between a real working scene and the
corresponding service for the actors doing the work.
Using the links, agents will offer suggestions
regarding best ways to carry out the current
activities that can be followed by different actors in
Figure 2. Hence, the database must provide a formal
way to describe various process activities and their
execution at a level of abstraction above the specific
instance of the content in which it is created.
5.2 Composition of the Database
In our research, a knowledge work process is
considered as a composition of a series of smaller
components, a unit of work (UOW). A UOW is
referred to any delimited piece of process activities
in enterprises, such as writing a report, making a
decision, designing a plan, etc.
A UOW is described using the WPSL we have
designed (Pan & Hawryszkiewycz, 2006b). The
activities of the UOW, their sequences, and the
relevant services for supporting these activities are
described through a particular architecture, which
includes the following six elements:
• Metadata that specifies commonly used entities
within the UOW, including its ID, Work goals,
and Preconditions;
• Role that illustrates various actors who may
participate in performing the UOW, for example
the ones shown in Figure 2;
• Service that describes all the supportive services
that may be provided for actors to support and
facilitate the completion of the UOW;
• Content that describes all the activities involved
in the UOW, each of which respectively
illustrates the type of the activity, what it does,
its outcomes and the corresponding services for
supporting the activity;
• Method that specifies all the possible methods to
evolve activity steps within the UOW; and
• Work plan that declares the links between a
method and its targeted actor’s category.
All these are compound elements and each
contains a set of more elementary elements. Some
are optional, and some may occur many times.
A UOW record describes the practices to evolve
the UOW by such a complex architecture. In this
way, the best practices of a knowledge work process
can be described by a series of such UOW records,
constructing the process guide database in Figure 3.
6 DEVELOPMENT OF THE
MULTI-AGENT
ARCHITECTURE TO REALIZE
THE SUPPORTIVE SERVICES
It can be seen from Figure 3 that the development of
the multi-agent architecture is another challenge to
improve knowledge work processes. The multi-
Suggestions on
the ways to
evolve process
steps which suit
the actor
Process guide
database
Working scene
The profile of
an actor
Figure 5: Role of the process guide database.
ICEIS 2007 - International Conference on Enterprise Information Systems
256

agent architecture is to realize the services that
support knowledge work processes. It will be
developed by extending the agent systems that
support business process management. Agents have
already been introduced into business process
management (Jennings et al., 1998). The problem is
most current systems focus only on enhancing well-
structured business processes where all the logical
paths can be defined in advanced. They pay little
attention to emergent knowledge work processes.
This research suggests agents be integrated with the
knowledge work processes to provide services that
not only assist participants to conduct structured
process routines but also facilitate the completion of
knowledge-intensive work, for example assisting
participating actors to identify a situation, assess
situations or define alternate courses of action, as
shown in Figure 2. It places a strong emphasis on the
guide for various knowledge-intensive activities.
The agents in the multi-agent architecture will be
designed to guide participants to advance process
steps based on the knowledge gained from the
process guide database and the actual process
progression. They will work at the background,
monitor the workspaces. When an event is detected,
they will use a Belief-Desire-Intention based
proactive reasoning to decide actions to respond the
event (Jennings et al., 1998). They will provide
supportive services for participants based on actual
work scene. For example, when the execution of a
process activity is being detected by the agents, they
will provide services to
• assist the actors to search, annotate, and make
sense of information required for the execution
of the activity;
• foster knowledge sharing among the actors
while performing the tasks specified in the
activity;
• encourage and scaffold the actors to actively
engage in the efforts for reasoning, problem-
solving, and decision-making using his or her
own personal preferred ways while taking a
responsibility in the activity;
• aid the actors to timely and accurately evaluate
outcomes of the activity.
A critical work for developing such a multi-agent
architecture is the development of the briefs,
intentions and reasoning strategies for the involved
agents so that they can collaboratively work for the
participating actors and provide them with right
services at right time to facilitate efficient
completion of the process.
7 OUTLOOK
The paper has presented an architectural framework
for supporting knowledge work processes and an
overall description on how the involved components
will be developed. The framework is currently under
development. It is hoped that the implementation of
the framework will enable
• process organizers to specify process activities
and recommend accomplishment strategies for
the participating actors by using the WPSL;
• process participants to get effective guide
provided by software agents. The guidance will
not only help them advance process steps but
also facilitate the increase of productivity and
the decrease of required costs;
• enterprises to get benefits from the improvement
of knowledge work processes.
REFERENCES
Davis, G. B. 2002. Anytime/Anyplace Computing and the
Future of Knowledge Work. Communications of the
ACM, 45(2). pp. 67-73.
Drucker, P. F. 1999. Knowledge Worker Productivity: The
Biggest Challenge. California Management Review,
41(2). pp. 79-94.
Hawryszkiewycz, I. 2006. Design Methods for
Collaborative Emergent Processes. BPM 2006
Workshops, INCS 4103. pp. 15-24.
Hawryszkiewycz, I. 2005. A Metamodel for Modeling
Collaborative Systems. Journal of Computer
Information Systems, 45(3). pp. 63-72.
Jennings, N., Sycara, K. & Wooldridge, M. 1998. A
Roadmap of Agent Research and Development.
Autonomous Agents and Multi-Agent System 1(1). pp.
7-38.
Markus, M.L., Majchrzak, A. & Casser, L. 2002. A
Design Theory for Systems that Support Emergent
Knowledge Processes. MIS Quarterly 26 (3). pp. 179-
212.
Mertins, K., Heisig, P. & Vorbeck, J. 2001. Knowledge
Management: Best Practices in Europe. Springer-
Verlag Berlin Heidelberg New York.
Pan, W. & Hawryszkiewycz, I. 2006a. Assisting Learners
to Dynamically Adjust Learning Processes by
Software Agents. International Journal of Intelligent
Information Technologies, 2(2). pp. 1-15.
Pan, W. & Hawryszkiewycz, I. 2006b. A Method for
Describing Knowledge Work Processes. in press.
Röll, M. 2004. Distributed KM - Improving Knowledge
Workers' Productivity and Organisational Knowledge
Sharing with Weblog-based Personal Publishing.
BlogTalk 2.0, “The European Conference on
Weblogs”, Vienna, Jul 5-6.
A FRAMEWORK FOR SUPPORTING KNOWLEDGE WORK PROCESSES
257