Journal of Knowledge Management Practice, August 2001

Support Systems For Knowledge Workers:

The Need For New Development Approaches

Marie Christine Roy, Laval University; Jean Falardeau  & Christian Pelletier, DMR Consultants Inc.

ABSTRACT:

Many organizations are facing the challenge of developing support systems for a class of employees described as “knowledge workers” (K-worker).  Because of the unstructured, novel and complex nature of the k-worker’s tasks, the traditional approaches for system development appear to be limited, even inadequate for these types of environments.  It is proposed that new approaches are needed to develop support systems for k-worker tasks.  In this paper, we will first describe some previous research work that provides support for the argument that the traditional methods are limited, even inadequate, for the reengineering of work processes and system development in the k-worker environment.  From this description, we derive three propositions related to the requirements of methods and approaches for these new work environments.  These propositions are later tested in three case studies of system development.

1. The Context Of The Knowledge Worker’s Job

1.1 The concept of Knowledge Worker

The knowledge worker concept derives its origin from the emergence of changes in the structure of jobs in organizations, this change being brought about by the higher complexity and the increase of information. Cushman et al (1999) and Harris (1999) define these workers as those who “gather, analyse, add value and communicate information to empower decision-making.  The nature of knowledge work is ad hoc, demand-driven and creative.”  Their jobs are best described as “decision cycles rather than a fixed sequence or series of well-defined tasks.”

The k-worker must therefore compose with an imprecise definition of the task to perform, its novelty and uniqueness, as well as participate in the formulation of a personalized and adapted response.  The tasks also require collaboration with other experts and information sharing.  The problems faced are most often complex  and unstructured, therefore requiring innovation and knowledge creation for their resolution.

This definition of a knowledge-worker diverges dramatically from the traditional definition of the information-worker, which has inspired the industrial economy of recent years.  Drucker (1994, 1999) mentions that the information-worker is considered as one that deals with a simple, well defined and sequential tasks.  The information requirements for the task are relatively simple and not necessarily up to date.  Furthermore problems have clear solutions because they are recurrent and well structured.  The organization around the information-worker does not encourage group work and is focussed on individual task description and responsibilities.

Given the task complexity of the k-worker, the support systems are critical for its accomplishment.  These support systems can significantly reduce the time required to accumulate and validate information, which is one of the k-worker’s main activities.  They should also provide support for solving complex and unstructured problems, in order to improve decision making effectiveness. These systems are generally classified under either decision support systems (Keen et al, 1978; Yee-man Fok et al, o1987) or performance support systems (Gery, 1991).

1.2 The need for new approaches for developing support systems for knowledge work

During the last 10 years, many organizations have restructured their activities, modified procedures and revised their business processes, in most cases to reduce operation costs and improve competitiveness.  These reengineering efforts were often accompanied, even leveraged,  by the development of new information systems.  These systems were integrated into organizational processes using analysis and design techniques of a traditional nature and were mainly developed to support structured tasks for the information worker.

Figure 1 : Traditional approach
 
Figure 1 illustrates the traditional approach of information system development.  In order to understand and analyze the actual work environment, the first step is to represent work processes in the organization.  Then, in order to specify the new working environment, the old processes are optimized and the role of new information systems are designed to improve the performance of these processes.  During the transition from the old to the new processes, the change agents will use a conventional approach for organizational transformation, where new tasks are defined while considering the resources and other constraints imposed by the work environment.  The development of the new information systems is performed using the waterfall model.

 

 

 

 

 

 


   

 

 

 

 

 

 

Organization development
 
 

 

 

 

 

 


Recently, there has been increased focus on knowledge management in organizations.  Zack (1999) mentions that organizations concentrate more on knowledge because they consider it as an important asset for survival.  To remain competitive, intellectual resources must be adequately managed.  The author also mentions that more effort is put on developing systems that capture, store and distribute knowledge.  Furthermore, more and more of the traditional information workers are becoming knowledge workers, in turn taking over the majority of the work force.  These changes are also affecting the systems needed to support these activities, which must be designed to address unstructured tasks and be linked to knowledge based systems.

The traditional methods described earlier are not well suited to develop these new types of systems.  For instance, Albers (1997, 1998) argues that the usual methods for task analysis are limited because their reliance on describing a sequence of actions is inappropriate for tasks that are ill defined and unstructured.  Figure 2 presents the general interrogations concerning the traditional approach when applied to k-worker support systems.  First, will a description of the business process be adequate to describe and understand tasks that are fuzzy or ill-defined?  The same interrogation applies to the representation of the new process.  As for the transition methods, do the usual ways to organize work still apply to the k-worker?  Similarly, are the traditional waterfall models for system development appropriate for these new work environments?

Figure 2: Limits in the traditional approach
 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


2. Understanding And Representing The Work Environment

As mentioned earlier, representing the work environment of the k-worker is far more complex than is the traditional business process  because it involves unstructured tasks in non-repetitive situations.  A number of techniques have been suggested by many different authors on the way to tackle the complexity and represent k-worker support needs.  In the following sections, we review these techniques in order to formulate a proposition as to the required approach for this phase.

The interview is the most common technique used to define information needs and to describe the knowledge acquisition process for problem solving (Moody et al, 1998).  The interview consists of questions asked directly to the experts of an area to collect information on the area and on the way they execute their tasks.  A more recently developed technique, the cognitive interview, is more focussed on triggering recollection of knowledge.  It is based on five principles, recreation of context, optimal concentration, and three principles that provide guidelines for triggering recall.  Moody et al. (1998) executed a study where they compared the effectiveness of standard interviews and cognitive interviews, and demonstrated that cognitive interviews were superior for knowledge acquisition activities.  This new technique showed more information rich and abundant recall, with more details related to the task of the expert’s domain.

Some other similar techniques have been described to conduct more in depth interviews.  For instance, Militello and Hutton (1998) provide an inventory of methods such as the simulation interview, which is based on providing decision making scenarios to elicit recollection, and knowledge audit, which results in an inventory of task-specific expertise.  In the same vain, Burge (1998) lists a large number of methods for eliciting knowledge in interviews, but also through other means such as observation, case study, protocol, role playing, etc.

Albers (1998) proposed goal driven task analysis to define information requirements for complex problem solving environments. This technique, although somewhat different from the cognitive interview, also puts emphasis on cognitive elicitation techniques.  In this case, Albers suggests using scenario development to define information requirements and group discussion to reconcile different points of view.

Although many methods have been developed to conduct interviews and elicit memory of information use and knowledge, few have suggested a way to represent this information to facilitate communication and understanding.  Militello and Hutton (1998) describe a technique called the cognitive demands table, which helps to detect conflicts, commonalities between tasks, and support the identification of new information that can be provided by the new system.  Albers (1998) recommends the use of a goal/information diagram to display possible paths to solving a problem, to bring out user needs and to reflect these needs in terms of user knowledge and desires.  It provides a graphical method of capturing the range of possibilities that a user can expect from a system.

Overall, the authors mentioned previously agree that for k-worker environments, the techniques and methods must be dedicated to a design that reflects the decision-maker’s mental model and the goals formed in the process of reaching a decision.  From this idea is derived our first proposition:

Proposition 1: An appropriate representation of the information needs of the k-worker requires cognitive-based techniques and scenarios.

An inappropriate representation of the user’s needs will lead to support systems that do not communicate information effectively, and therefore to low satisfaction and low impact on decision making performance.

3. Representing The Target Work Environment

The techniques presented in the previous section seem to be well adapted to understand the k-worker’s environment and task.  Of course, this first step in the analysis of the k-worker may provide some important insights onto solutions and transformations to support their tasks and the type of information systems required.  However, a second step requires that solutions be designed and tested with future users, so as to identify the most appropriate or optimal designs, to achieve consensus if there are many users and to provide a platform for user learning and experimentation. 

Prototyping has often been suggested as an approach well suited for the development of decision support systems (Cerpa, 1996; Turban & Aronson, 1998).  A prototype can be a variety of things, from a paper mock-up of computer screens to a completely functional system, but its main objective is to illustrate how the future system will be designed and used.  When prototypes are used as a system development approach, they can either be used very early in the process (to better understand the work environment, for instance) or at a later time, when the system’s specifications are well defined.  They can also be used to explore different design alternatives (by using many prototypes for the choosing), or only used for the fine tuning of system functions.  They can be expendable and cheap to produce, or evolve to become the final system  (Hardgrave et al , 1999). 

Although prototyping is often described as a system development method in and of itself, the ways in which the prototype is used will vary greatly from one development project to the next.  However, the general objectives that are pursued by the prototyping approach are better served when the prototype is used earlier in the process.  These objectives are 1) to clarify user needs and system requirements, 2) to test high-innovation projects, 3) to increase interaction and communication between users and developers 4) to increase user learning (Hardgrave et al, 1999; Turban & Aronson, 1998).   As opposed to traditional approaches, where the objective is to optimize a process that is well understood by the development team, an iterative prototyping approach relies more on trial and error, evolutionary construction and user-developer teamwork.

In the context of developing systems for the k-worker, there will be high emphasis on intensive user involvement and the testing of different solutions, because of the nature of the work environment.  The prototyping approach will be well suited for this setting.  We therefore state the following proposition:

Proposition 2: An iterative approach based on prototyping will be required to obtain an adequate design of k-worker support systems.

4. The Transition From The Old To The New Work Environment

The appropriate management of the transition from the old to the new work environment is crucial for the successful implementation of the new information system.  The change is delicate to handle because it must take into account all the human factors that interact in the organization and that are impacted by the new system modalities.  In the context of knowledge work, users have a high level of education, expertise and occupy upper levels in the organizational hierarchy.  These workers are constantly in an information search mode and use communication networks that they have developed to find the relevant information.  Very often, they have established their organizational standing through social and political means, and these contingencies must be considered when integrating the proposed changes in their work environments (Pasmore, 1988; Stebbins & Shani, 1995).

As mentioned by authors such as Yeeman Fok et al. (1987) and Fox (1995), a socio-technical approach attempts to provide a broader view than the one typically adopted in traditional approaches.  It advocates the development of information systems that considers the joint optimization of social and technical systems for optimal organizational performance.  Some recent studies have attempted to show how socio-technical systems theory could be adapted to accommodate non-linear throughput and knowledge work (Pasmore, 1988; Stebbins & Shani, 1995).

To study the social system, the authors recommend a high level of user involvement in the design process and the use of deliberations on key issues.  Results of the deliberation analysis are used for the redesign or defining the ideal organization, aligning knowledge with influence in decision making and identifying improved formal structures for integration of knowledge work.  We therefore state the following proposition:

Proposition 3: A successful implementation of a new organizational design, accompanied by new information systems for k-workers requires high involvement and deliberation of k-workers on social and political aspects in the environment.   

5. Case Studies

In order to provide support for the three propositions stated earlier, we performed three case studies in organizations where support systems were either developed or in the process of being developed for knowledge workers. 

5.1 Case #1: Ice monitoring system for the Canadian Coast Guard

Our first case study was performed at the Canadian Coast Guard.  During the winter months, the Saint-Laurence River freezes up and a navigation route must constantly be maintained free of ice jams for commercial shipping.  Monitoring operations were typically performed by helicopter or through ship reports, information was made available to the experts through telephone, fax or radio.  In 1994,  an ice jam on the river halted shipping for more than a month and generated flooding that led to 30 million dollars in lost revenues and damage.  This event led to the development of a sophisticated monitoring system based on camera, radar and sonar surveillance, and this information was available from a desktop computer to 3-4 experts.  Although the system is still being fine tuned and improved, it was fully operational in 1999 and has had a significant impact on surveillance costs and maintenance of the maritime route.

Because the task of river surveillance in winter conditions requires decision making that is highly unstructured and a high level of expertise, the person performing this task can be classified as a typical knowledge worker.  The new system provides most of the information and support for performing this task.

In order to collect information on the development of this monitoring system, we interviewed the project leader, whose main responsibilities was to develop and maintain the new system, and the most important user, an maritime expert who has been responsible for the winter surveillance team since 1994.

5.1.1 Methods used to represent the information needs of the knowledge worker

In this case, there was no specific method or phase to define the needs for the new system other than informal exchanges between the developer and the main user.  However, the developer had been working with the coast guard for many years and was in a close working relationship with the user.  Therefore the developer described himself as very knowledgeable about the needs of the user at the beginning of the project.  No formal description of the decision process was attempted in this case.

5.1.2 Methods used to develop the system

Many different system experts were involved in the system development, mainly because of the highly sophisticated technology required.  There were specialists in computer programming, image processing, hydraulics and sonar measurement participating in the development team to build the infrastructure needed for collecting the required information along the Saint-Lawrence River.

Specifications for the user interface were developed through prototypes and successive interactions with the user.  Initially, the screen layout was presented to the users and modified, then an operational prototype was available for testing.  Although the initial trials were unsatisfactory due to the unreliability of the communication channels and the monitoring equipment, the system still presented major improvements for monitoring effectiveness. The system has been constantly upgraded since the beta version and is now described as being fully functional and responsible for one of the most reliable ice surveillance operations in the world.

5.1.3 Transition from the old to the new systems

The new systems was implemented through many iterations and was therefore gradually integrated into the expert’s job.  The user describes the system as providing an easier environment to perform the surveillance task, therefore this facilitated its acceptance and use.  Since most of the system use is performed by one expert, there were no difficulties related to the social environment.

5.1.4 Implications and conclusion

Although this case describes de development of a system to support knowledge workers, it has two characteristics that made it more simple to tackle than is typical:  It has just a few users, which facilitates the process of defining the user’s needs and involving the users in the system development phases.  Also the necessity to understand the social and political environment is much less critical for successful implementation.  The task is performed by highly qualified experts, which means that the system need only provide the basic information for decision making and not the additional knowledge necessary to process the information.  The case shows how constant user involvement and iterative prototyping led to a highly satisfactory surveillance system.

5.2 Case #2: Support system for Client-server architects at Fujitsu

Fujitsu had initially developed a system where client-server architects would find a knowledge base necessary to perform their work.  Although all the necessary knowledge was available in the system, it was not structured in a way to support a specific task and provide guidance as to the relevant knowledge to use in the task.  The system was therefore not used and required some major changes to be more task oriented.  A new system development process was undertaken to analyse the knowledge worker’s task and to construct a more effective system interface.

In order to better understand the development process, we interviewed the project leader for the new system.  Since this new system was never implemented because of budget cuts, it was not possible to evaluate the success of the system nor to interview a user. 

5..2.1 Methods used to represent the information needs of the knowledge worker

To improve the system’s fit to an architect’s task, the decision process needed to be well understood by developers.  Many different methods were used to document this process:  Day long meetings were held with groups of architects for them to express their needs and make compromises.  Observation of the architect’s work environment and decision process and interviews were also performed to document the process.  Cognitive mapping was used to represent the decision making situations, including both implicit and explicit knowledge. 

5.2.2 Methods used to develop the system

The development team included many different specialists:  Programmers, analysts, cognitive mapping specialists, ergonomics specialists, human-computer interaction  and communication experts.   The process required some new approaches and a new system development method was defined as the work evolved.  Major difficulties were encountered in trying to understand the knowledge worker’s environment and to integrate task knowledge into the new system.

5.2.3 Methods used to perform the transition from the old to the new system

The new system was not implemented.

5.2.4 Implications and conclusion

Compared to the coast guard surveillance system, in this situation there were many users, many of these were not experts and required the system provide some guidance as to how to use the available knowledge in specific situations.  Therefore the design of the system necessitated that the development team analyse the decision making situations with new representation techniques that would not only capture the knowledge needed but also used to facilitate communication and deliberation among users.  Although the system was not implemented, users perceived that the new approaches were to offer more promise as to the quality and usefulness of the resulting system.

5.3 Case #3: Analysis of insurance coverage by brokers

The last case study was performed in a large insurance firm which offers individual coverage for life, invalidity, travel, medication, health and injury.    The insurance industry is evolving towards world wide competition and greater diversification of its products.  The insurance broker, although still performing an individualized, face-to-face approach to analyse insurance needs and to provide recommendations, requires more sophisticated tools to access relevant information and to record data on visits and sales.

The insurance brokers are equipped with portable computers which contain databases of statistics and other information to support the sales process.  These computers also provide questionnaires to guide the broker as to the relevant information to collect from the customer.  However, no support was provided to suggest the appropriate insurance packages, based on the customer profile.  The sales process is highly dependant on the broker’s expertise and the customer very often cannot understand the link between his or her situation and the insurance coverage suggested by the broker.  Because of the greater variety of more personalized packages offered in recent years,  the profile of the broker is changing from a typical sales person to an expert financial planer. New tools are needed to provide adequate support for the broker’s job.  The company therefore decided to undertake a large development project to implement these tools on the broker’s portable computer.

5.3.1 Methods used to represent the information needs of the knowledge worker

When initially developing the new system for the insurance brokers, the analysts faced three important roadblocks.  First, the solutions were based on a detailed analysis of the sales process and its related activities.  The first phase of this process, acquiring the customer’s confidence, was described and supported as a linear set of steps through which the broker had to go through.  The first prototype was considered too rigid and inadequate to support what was a more individualized and non-linear activity.  Also, the system required that the broker enter predefined information which the customer was not always willing or able to provide.

A second problem was related to the standardisation of the process.  The system required a long series of actions for very simple situations, but on the other hand did not adequately support more complex situations.  Finally the third problem was that the initial solution didn’t include the knowledge or competencies of the brokers.  This knowledge could not only permit more effective support for experienced people, but also be important to guide novices on how to execute the sales process.  Analysts, developers, process experts and brokers alike concluded that they should change their approach in order to solve the problems described above.

The first solution was to involve insurance experts; one broker and an actuary were added to the development team.  Then the focus was put on understanding the broker’s task and many different scenarios and situations were analyzed.  Appropriate support was defined based on each case.  

5.3.2 Methods used to develop the system

The development team abandoned the use of a process oriented formalism to represent the sales process and began documenting the problem solving nature and needs in each of the identified sales situations.  The results of this analysis were incorporated into a new prototype. The proposed solution permitted more flexible support and more extensive analysis and diagnostic functions.

5.3.3 Methods used to perform the transition from the old to the new system

The new system is still under development.

5.3.4 Implications and conclusion

The evolution of insurance products available to consumers from a few different packages to a wide variety of insurance and financial products has led to a more complex and knowledge intensive sales job.  The broker can therefore be considered as a knowledge worker. 

The case shows that the analysis of the knowledge worker’s job could not be described as a predefined process with a fixed set of activities, but a complex web of rules and knowledge that vary depending on the customer’s situation.  The involvement of experts in these decision making areas was found to be critical for the development of appropriate support solutions.  Furthermore, it became apparent that the traditional tools used to represent a task (process analysis formalisms) were inadequate because of its highly flexible and iterative nature.  Although the system is still in the early phases of development, a readjustment of approaches led to a prototype that offers more potential for effectively supporting the brokers in their job.

6- General Discussion

Table 1 provides an overview of the results observed in the three case studies.  Our first proposition stated that to have an accurate and useful understanding of the k-worker’s job, representations needed to be based on cognitive techniques and scenarios.  In cases 2 and 3, we found that the use of more traditional process oriented approaches led to the development of solutions that were inadequate.   In both organizations, a new analysis was performed with user involvement, scenarios and cognitive mapping, techniques which seemed to offer more promise in capturing the task characteristics and requirements.   In the first case, there was no formal method used, either of a traditional nature or otherwise.  This was mostly due to the fact that the developer was very familiar with the task and in a very close working relationship with the main user.  These cases therefore offer some support for our first proposition.

The second proposition stated that the design of new support systems for k-workers has to rely extensively on prototyping, in order to increase user involvement and improve the quality of the proposed solutions.  In the three case studies we found prototyping to be critical to the development efforts and used early on in the process.  In case 1, many different prototypes were used, the last one having evolved into the new system.  In cases 2 and 3, prototyping was used to validate some proposed solutions.  However, since both these projects were only observed in early phases, it was not possible to know how these attempts were important to implement successful support systems.  Nonetheless, it seems quite evident from these cases that prototyping was central to developing design alternatives, therefore also providing some support for proposition 2. 

In our third  proposition, we stated that the successful implementation of K-worker support systems requires high involvement and deliberation of k-workers on social and political issues.  Our case studies did not provide much information on this aspect since two of the three had not implement their systems.  Furthermore,  the few number of users in the first case reduced the need for deliberation between them.  However, we generally observed the importance of the high involvement of users in the three cases.

We stated three propositions related to important steps in systems development and organizational change in the k-worker environment.  The case studies provided some support for these propositions.  Figure 3 illustrates the overall approach to k-worker support systems development.

Table 1: Summary of the three case studies

 

Case 1

Coast Guard

Case 2

Fujitsu

Case 3

Insurance

Information needs representation

Few users

No formal definition of decision process

Meetings

Observation

Cognitive mapping

Interviews

First: Process analysis

Second: Scenario and user involvement

System development methods

Successive iterations

Prototyping

High user involvement

Multi-expert team

Knowledge integration in the system

Prototyping

Prototyping

Sales situations and scenarios

Transition from the old to the new system

Gradual implementation with incremental changes

Not implemented

Not yet implemented

 

 

Figure 3 : Proposed approach
 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


 

 

Researchers and system developers are just beginning to discover how to provide support for knowledge workers and how to go about approaching the development of these systems.  Some innovative experiences described in our case studies seem to show that traditional methods for process analysis and development fall short in providing the tools needed for the particularities of these environments.  In two of these cases, the development team radically changed their way of approaching the situations because of the shortcomings of their usual way of developing information systems.

The case studies described in this paper provide a very limited insight on the ideal approach to k-worker support system development.  More research is needed to ground the propositions stated here into generalizable recommendations.  For instance, future studies should be performed on a larger sample to extract the success factors in the implementation of these types of systems.  Also, researchers and developers must also work at defining and training themselves on the new techniques that are required for these new environments.

7. References

Albers, M.J., Goal-driven task analysis: improving situation awareness for complex problem-solving, Proceedings of the 16th annual international conference on computer documentation, http://www.english.ttu.edu/grad/albers/vita/sigdoc98.htm, , 1998

Albers, M.J., Information Engineering: Creating an Integrated Interface Poceedings of the 7th international conference on human-computer interaction http://www.english.ttu.edu/grad/albers/vita/hci97.htm, , 1997.

Albers, M.J., Information design considerations for improving situation awareness in complex problem‑solving, Proceedings of the 17th annual international conference on computer documentation, 1999.

Burge, J.E., Knowledge Elicitation Tool Classification.  Artificial Intelligence   Research Group, Worcester Polytechnic Institute, http://www.cs.wpi.edu/~jburge/thesis/kematrix.html, , 1998.

Cerpa, N., Verne, J., Prototyping: some new results, Information and software technology, vol. 38, p. 743-755, , 1996

Cushman, A., Fleming, M., Rosser, B., Hunter, R., Harris, K., The knowledge management scenario: trends and directions for 1998-2003, http://gartner6.gartnerweb.com/ggbin/ggtoc, 1999.

Drucker, P.F., The age of social transformation ,  http://www.theatlantic.com/atlantic/politics/ecbig/soctrans.htm, 1994.

Drucker, P.F., Knowledge worker productivity: the biggest challenge, California Management Review Reprint Series, Volume 41, Number 2, 1999

Fox, W.M., Sociotechnical system principles and guidelines: past and present, Journal of applied behavorial science, Vol. 31, No. 1, pp. 91-105, mars 1995

Gery, G., Electronic performance support systems: How and why to remake the workplace through the strategic application of technology.  Boston, MA: Weingarten Publications, 1991.

Hardgrave, B.C., Wilson, R.I., Eastman, K. Toward a contingency model for selecting an information system prototyping strategy.  Journal of Management Information Systems, vol. 16, no. 2, 1999

Harris, K., So, you’re a knowledge worker: what do you do all day, http://gartner6.gartnerweb.com/ggbin/ggtoc , 1999.

Havens, C., Enter, the chief knowledge officer, CIO Canada, p.36-42, 1996

Keen, P.G.W., Scott Morton, M.S., Decision Support Systems: An Organizational Perspective Reading, MA: Addison-Wesley, Inc. , 1978

Militello, L.G., Hutton, R.J.G., Applied cognitive task analysis: a practitioner's toolkit for understanding cognitive task demands. Ergonomics, vol. 41, no 11, 1618-1641, 1998.

Moody, J.W., Blanton, J.E., Cheney, P.H., A theoretically grounded approach to assist memory recall during information requirements determination. Journal of Management Information Systems, Armonk, 1998

Pasmore, W.A., Designing Effective Organizations: The Sociotechnical Systems Perspective, Wiley, 1988.

Stebbins, M.W., Shani, A.B., Organization design and the knowledge worker, Leadership & organization development journal, Vol. 16, No. 1, 1995

Teng, J.T.C, Grover, V., Fiedler, K.D., Business process reengineering: charting a strategic path for the information age.  California Management Review, Berkeley, 1994

Turban, E., Aronson, J.E., Decision Support Systems and Intelligent systems, Prentice Hall, New Jersey. 1998.

Yee-man Fok, L., Kumar, K., Wood-Harper, T., Methodologies for socio-technical-sytems (STS) development: a comparison, ICIS Conference Proceedings,, pp. 319-334, 1987

Zack, M.H., Developing a knowledge strategy.  California Management Review; Berkeley, 1999

 

The authors may be contacted at Marie-Christine Roy, Université Laval, Ste-Foy, Québec, CANADA G1K 7P4; by email at Christine.Roy@fsa.ulaval.ca; by telephone at (418) 656-2131; by fax at (418) 656-2624