Albert F. Bokma
Centre for Electronic Commerce, School of Computing, Engineering and Technology, University of Sunderland, St. Peter?s Campus, Sunderland SR6 0DD, United Kingdom, Tel.: +44 191 515 3233, Albert.Bokma@sunderland.ac.uk
The management of corporate knowledge is becoming highly relevant in the light of increased competitive pressures, where organisations have to operate extremely effectively and efficiently in order to survive. The industry of the leading economic countries is increasingly forced to focus on innovation and superior technology, rather than mass-production capabilities, and the corporate knowledge and skills base, rather than production power, is becoming the lifeblood of successful companies. Apart from the commercial applications that particularly highlight the relevance of such solutions, it is also self-evident that these solutions are also important for individuals in a networked, information society age; with an almost exponential increase in both individuals that are connected in networks and a similar increase in information few understand how to adequately manage and control.
The adequate management of corporate knowledge is becoming crucial and represents a significant factor of growth and will become an important factor to sustainable growth of the enterprise for the future. This has been highlighted by [Sveiby 92] where he suggests that among the fastest growing companies in the US about half are knowledge intensive organisations. Although this includes mainly companies that from the services sector (also notably consultancies such as Andersen [Elliott 97], Ernst/Young [Emerson 97] or IT companies like IBM [Huang 97]), it stands to reason that the industrial sector where the products encapsulate a great deal of know-how will be no different [Apostolou et al. 98]. In the automotive sector for instance, companies are now increasingly looking into how to best manage the skill-base they have and how to maximise their effective input into product design and production. Thus, for instance, SAAB of Sweden and their CONSENSUS project to try to find and exploit new ways of managing the complexities of design and production processes including the management of flexible design teams and there are others such as BMW and Caterpillar.
The use of flexible design teams in the agile enterprise, for example, has definite advantages in allowing well-focused interdisciplinary teams to collaboratively solve design tasks. The problem with this approach is however the risk of loosing valuable design knowledge when the teams disband and members join other teams on other tasks. Though in this case some positive "cross-pollination" happens as the team members bring more experience to the new team and new task, oftentimes the knowledge and expertise generated in the old team, in its totality, is multi-disciplinary in nature, and no individual member may necessarily be able to effectively use those knowledge components that are outside their subject-area. In addition, there is also the risk that other teams on other, but similar design tasks may "reinvent the wheel" and duplicate efforts made elsewhere, thus giving rise to inefficiencies. What would therefore be required in this particular example, is a way of recording not just the design, the rationale and expertise and the people from whom that expertise originated and on what it was based, and to subsequently make that accessible to others in the organisation in order to facilitate the reuse and avoid redundancies and repetition.
The example of flexible design teams is, however, not much different from other aspects of collaboration found elsewhere, where there are good reasons for the recording and reuse of the often tacit knowledge and expertise generated by members of the organisation. In an age of exponential information growth the management of that information both for the organisation and even for the individual is becoming increasingly crucial. A number of solutions have been developed including a plethora of commercially available products, but most of these either provide solutions for well-structured domains which lend themselves to systematic knowledge capture and reuse or solutions are tailored for a particular application and enterprise but not so easily transferable to a new situation. The focus of CogNet is on situations which are not easily standardisable, either in terms of processes or categorisations and where a flexible approach is required. One such area is the management of innovation and the creation of creative environments. Oftentimes the emphasis is on the isolation and reuse of knowledge per se, without attaching a great deal of importance to the sources of knowledge i.e. the people involved; CogNet by contrast focuses on these links as an important aspect of the solution especially when knowledge management solutions in an enterprise are not only to serve as repositories of useful information to be consulted as appropriate but as a means of effective communication.
In the following sections we look at the issues to be addressed and references to current work in this area. In the subsequent section we outline the CogNet approach we propose for its solution as well as giving a typical scenario in which it is expected to fit, before drawing relevant conclusions.
2. Key Issues and Related Developments
In order to address the problem of knowledge management adequately in an organisation requires in its totality not only a technical solution, but needs to be accompanied by an analysis of the current situation and future requirements of an organisation. This should serve to establish a set of organisational goals, strategies and measures to guide the organisation through the process of adopting an effective knowledge management implementation. Thus a business oriented approach will have to complement a technical solution. The present paper acknowledges the need for such a complementary measure, and which is assumed, but the details of which will be discussed elsewhere and are, for the moment, excluded from our considerations. As far as the technical solution is concerned a number of aspects need to be covered and which fall under the headings of the standard knowledge life-cycle:
A number of approaches and methods have been proposed, but not all of these concern the present investigation directly as they concentrate on business and management issues, such as [Askari et al. 95], without going into detail about the technical ICT implementation and its functionalities. Amongst examples of methods that are more closely related are [Davenport et al. 98] and [Mentzas et al. 98]. These propose a systematic approach to the knowledge management analysis process, its design and specification, as well as building on the theories of [Nonaka et al. 95] for the dynamics of a knowledge-creating company, focusing on the dynamics of knowledge generated in various contexts and how to manage it. Likewise, another interesting development is [Sveiby 97] proposing a method for the management and measurements of knowledge-based assets. There are also interesting developments in the use of conceptual structures based on [Sowa 84] such as [Thanitsukkarn et al. 98] and others and there also exists work on knowledge management in distributed and dispersed environments [Heijst et al. 96] or work on knowledge management and review [Grundstein 95]. Knowledge management for effective communication and active documentation in the enterprise [Durstewitz 94] is interesting in this respect as well as the need for multidisciplinary support [Kühn et al. 97]
There are various approaches that have also been proposed for automated knowledge acquisition and data mining that will be relevant for the purpose of harvesting pieces of knowledge that already exist explicitly or implicitly in existing Information Systems within the organisation. This area will be examined in more detail in another place, and, for the purposes of this paper, we shall concentrate on the knowledge organisation and management issues where the individuals will be involved directly and who will themselves make explicit contributions to the corporate knowledge structure and repositories. In order to preserve and reuse existing expertise and knowledge and to manage an increasing body of information, there is a need for a mechanism, that can effectively record knowledge in a form that is useful to the individual and can easily be accessed by others. The facilities and capabilities are therefore important both from the perspective of the individual and the corporation; the tools have to be extremely useful in order to make it attractive for the individual to use to record their knowledge and hence facilitate their work. It also has to be organised in a standardisable or normalisable fashion to allow interested parties to effectively search through the entire skill and knowledge base of the corporation to locate appropriate knowledge and people for the task at hand.
The knowledge organisation problem is at the heart of the investigation and there are a number of related areas of covering knowledge representation as well as access issues. A great deal of knowledge management projects that have been implemented in organisations have been based on specific management consultant input and analysis to determine the needs of the specific enterprise and then implemented with a number of standard technologies. These include access and maintenance of information repositories (distributed databases) by a distributed workforce together with communication mechanisms (e-mail, workflow and productivity tools) and dedicated knowledge guardians (personnel) together with knowledge sharing organisational approaches. There are a number of examples specifically from the service sector [Elliott 97, Emerson 97 etc.]. Our focus will be on a single, generic approach and method, which will be supported in future by a suitable tool and although lessons are to be learnt from these experiences they represent specific approaches and custom solutions and are therefore only partially relevant. As touched on briefly in the previous section, we propose a semantic approach to corporate knowledge management and there are consequently two areas of current research that are clearly relevant, namely work on ontologies [Farquhar et al. 95] and work on semantic/conceptual approaches [Thanitsukkarn 98]; likewise on solutions for distributed personnel in collaborative settings [O?Leary 97], [Tennison et al. 98] and [Dominique 98]
As we have stated earlier, it is our firm conviction that knowledge cannot meaningfully be seen in isolation to the people involved. Hence our fundamental premise is that the knowledge captured, represented and reused needs to be connected to the beholder and that, if implemented appropriately, this has the added benefit that it can also connect individuals as well as giving individuals access to the knowledge. Consequently, the representation which will be at the heart of the approach needs to accommodate and interconnect the following aspects:
3. The CogNet Approach
It is our contention that knowledge and skills within an organisation do not always easily fall within simple categories and follow a standard pattern. Therefore a solution is required which is flexible to record a variety of items. In addition knowledge should not be seen in isolation but should be connected to the individuals that are the source of that knowledge as well as the context in which it is to be interpreted. A connection will therefore need to be made between individuals, the items they have knowledge/skills for and the specific knowledge/skills themselves. In addition, it is usually counterproductive to view such knowledge in tabular form which is convenient for the computer but not for the individual who wants to explore it and more intuitive and meaningful ways of presenting and navigating will have to be found. The proposed solution will use a semantic approach in the form of a conceptual graph, combined with a suitable ontology that allows to make the relevant links and record additional information. This can then also be the basis for the eventual use of intelligent techniques to search and highlight implicit knowledge that is not recorded prima faciae.
The approach that is proposed will use in essence a conceptual net as opposed to a semantic nets in that the latter is more fine grained operating at a natural language level and how natural language is used to denote and describe reality in general or in a restricted sense. Semantic nets have per se a great deal of beauty and appeal, however for most practical applications they are too elaborate and unmanageable (at least at present). Instead a conceptual approach is proposed and which is more coarse-grained and represents more complex concepts and operates with a standard ontology suitable for the problem domain at hand. At the same time the basic structure will be expandable by the user to add new categories and classifications to suit their requirements. The conceptual net will also have a number of dimensions and which allows for different access strategies to quickly and easily trace pieces of knowledge and the people associated with that knowledge.
A sophisticated access mechanism will allow to manage the access of people in the organisation, to enable individuals to record their contributions to the corporate knowledge structure and restrict access to certain pieces of information not designed for a wider circulation. In addition it will allow to monitor and record access, so that the "consumption" of the knowledge and expertise can be traced and acknowledged. Apart from providing the facilities to allow the easy recording of knowledge and expertise and its tracing and reuse, it is important to provide facilities that generate an incentive for its use and to encourage individuals to contribute to it and make use of it. Thus the access management functions are designed to record individual contributions and to acknowledge access to these contributions and their reuse. Thus the added value generated by individuals can be assessed and rewarded.
The fundamental assumption of the CogNet project is that in developing an ultimately successful knowledge management solution for enterprises, it is important to take a holistic approach and give equal weight to organisational and technical aspects. In the following, the CogNet approach is specified, starting with an explanation of the Individual Knowledge Organiser that will connect with other such organisers and an Enterprise Cognitive Network
3.2 Individual Knowledge Organiser
Most people will be unable to keep full control and knowledge even over all the document files created by them or stored by them on their own computer. Given that the current proliferation of information is unlikely to level off, but will be a fact of life, there is a great need for individuals to keep track of both information and knowledge. By this distinction we mean the documents created or imported by them and the knowledge and views they have both about the documents in question and any other referents. As stated earlier knowledge is not to be seen in isolation but needs to make reference to the context it is placed in and documents it is connected to. Typically individuals hold information in a computer in files in a hierarchical file system designed to keep some order among documents and in a set of folders. This is essentially an electronic version of what otherwise would have been a physical system of filing cabinets, folders and files in these folders according to some classification system. The aim is to replace this approach with an individual knowledge organiser (IKO) that allows in the first instance to:
Likewise the user will be able to manage other categories of information such as contacts, events, products, processes etc. either by storage in the ontology, or through the ontology pointing to their physical location on the user?s computer or to remote resources. The meta information in each case being a unstructured text field (a separate document if large and complex) can therefore be attached to the ontology and relocated by its classification. The meta knowledge will then be used to record thoughts, comments and knowledge in an explicit form and can be interpreted within its context by examining the links emanating from it and that point to its referents as well as the link from the ontology pointing to it. The Figure 1 below shows this in a summative form without explicitly depicting the ontology and which is contained within the thought bubble (in addition to the ideas and knowledge indicated). Indicated are links (from left to right) to documents, contacts, events and products, as well as personal qualifications and skills (bottom left).
Figure 2 depicts the ontology in a simplified form. The "known universe" depicted at the top is subdivided into a number of key categories and which are in turn subdivided into sub-categories and so on by way of a standard classification for an enterprise but which can be customised and extended by the user. By contrast figure 3 below depicts the relation and access potential between a document and the user and which is directly linked to its author, but which can also be reached through thematic navigation of the knowledge, ideas and thinking recorded in the Personal Knowledge Organiser.
In order to collect the meta-information from the user and to correctly insert the documents in the ontology framework the user will be prompted and be able to use pull-down options to quickly achieve the correct classification and to record and which will be automatically indexed by relevant keywords. These functions will be user customisable. The ontology will be implemented in a conceptual graph and which the user will navigate graphically or in tabular form to allow rapid browsing and searching. The Personal Knowledge Organiser will operate on the user?s computer independently in standalone mode, but which will connect to the Enterprise Cognitive Network Server when network connections are available. In that event the user will have access to central resources and (given appropriate permission) to explore Personal Knowledge Organisers of colleagues. This is described in more detail in the following section.
3.3 Enterprise Cognitive Network Server
The aim of the CogNet project is to develop a Creative Knowledge Environment as a solution candidate for new knowledge mediation processes and virtual knowledge management. In order to support these processes of creating innovation out of knowledge at an enterprise level, the project will develop a creative knowledge sharing environment. Using sociological and organisational knowledge on learning the project focuses to shape technology overcoming the logistics perspective (integration and standardisation) and make human intelligence the guideline for the design of our solution. Using the results of "group dynamics" and "organisational learning" [Nonaka et al. 95] and others, the project will develop a creative environment for a variety of types of creative activities and motivate enterprise innovation (introducing new roles as context manager and knowledge broker).
The results will be of particular relevance for situations that are inherently unstructured and difficult to "standardise", such as situation where creative activity is required such as in innovation management in the enterprise, but the results should, in principle, be equally applicable to other aspects of the business. The approach will be expandable to the enterprise as a whole and, in terms of a stakeholder approach, be capable of supporting also the meaningful interaction between different interest groups such as management, employees, suppliers, customers, investors and the public.
What is in essence required to fulfil the goals of corporate knowledge management in this wider sense can be summed up under three distinct headings:
Once Individual Knowledge Organisers have been populated by their users it is possible to share knowledge and information with others by connecting the organisers and thus enabling a horizontal flow of knowledge depicted by the oblong box. Thus a user can either proceed to remotely view the shared knowledge structure and content of his/her colleague in isolation or to superimpose the two in order to be able to have rapid and easy access to either. As described in the previous section it is envisaged to make the private knowledge ontologies accessible both in tabular and graphic form to allow alternative navigation and browsing access. In either case it is envisaged to allow to switch the view to distinguish between knowledge ontologies from different users through the use of colour. As stated previously not all knowledge represented will be sharable and therefore only selective access will be granted as defined by the users or organisational requirements.
The CogNet system should free the user from viewing and organising information on their PC by directory hierarchies and provide access to central information and knowledge resources through the Individual Knowledge Manager (IKO) described in the previous section. IKOs from different will link through a special intranet that will run the Enterprise Cognitive Network Server (ECON). More details about the implementation are given in the following section and the present section gives an overview of the approach and the rationale behind it.
In order to enable sharing the knowledge encapsulated in the IKO of a particular user there needs to be a link to their IKO. The CogNet system will link individual knowledge organisers by a horizontal connection in the corporate systems to allow visibility and flow of individual knowledge and information among colleagues, as indicated by the box in the diagram above. In the CogNet framework the user will be able to start with any of the available categories and then to explore the network in any direction by traversing the links associated with the item under inspection. They will be able to freely move to other concepts or explore any items connected with a given concept. For example given that the user has taken an interest in a particular document he/she will be able to explore any authors associated with it as well as any other individuals that have made a reference to the item and offered their views on it. Likewise if the document is associated with a production process and design for instance the relevant documentation will also be accessible and their association be visible.
The architecture will use a conceptual net to create an interconnected network making the relevant links and enabling the recording of additional information. The conceptual net will also have a number of dimensions allowing for different access strategies to quickly and easily trace pieces of knowledge and the people associated with that knowledge. A sophisticated access mechanism will allow for the management of the accessibility to the variety of knowledge residing within the people in the organisation. By enabling individuals to record their contributions to the enterprise knowledge structure and by designating access will allow for monitoring and recording the access in a way that the variety of knowledge and expertise can be exploited, realising a virtual innovative knowledge environment. Apart from providing the facilities to allow the easy recording of knowledge and expertise and it?s tracing and reuse.
Figure 5, above, shows a simplified representation of the CogNet concept and the full interconnection between the main categories in the cognitive enterprise network and which can be navigated from any point. The lines indicate the interconnections between different categories and viewpoints. A number of individuals are shown, each with their own cognitive perceptions on ideas and knowledge of categories such as people, documents, products, events, etc.
The user will have a navigation tool that allows the graphical exploration of the conceptual net to access items and explore the connections associated with them. Nodes will represent concepts and items such as documents, contacts, designs etc. and which are interconnected by links of different type indicating such relations as ownership, comment, reference etc. In addition, the user will be able to selectively switch on/off specific categories from the view in order to facilitate navigation and exploration. The diagram below depicts a view from a particular document in the centre and with other items pointing to and from it:
The CogNet concept described so far has concentrated on the information and associated knowledge and how this works in terms of the Individual Knowledge Organiser and the Enterprise Cognitive Network. These structures enable the individual and the enterprise to record both pointers to interesting information as well as knowledge associated with it and thus to enable access both to the items of knowledge and the individuals associated with it. In order to facilitate the creative activity that precedes innovation and that drives the development of the expertise and product features and product range there is a need for environments in which the creative interaction between individuals can happen. Creative activity requires a great deal of freedom, yet a focus is required in order to facilitate a rich and meaningful exchange of ideas. In this way likeminded individuals can brainstorm, criticise and make constructive suggestions and exchange them with their colleagues. These ideas, once tested, can then feed into the product development process. In figure 5 the category at the bottom points to Fora and which are the focus for creative activity. These fora will have a theme and where individuals can join and discuss with their colleagues but rather than establish endless straight lists as in newsgroups to add to create and add components to a cognitive structure as used elsewhere in CogNet. This will allow the users to publish position papers and link them to existing contributions by way of specific types of links to denote referencing, criticising, approving disapproving and the like. These fora will also be moderated by a context manager (a designated person responsible for that forum) that will moderate the evolution of the discussion to avoid discussions to become stale and unproductive. As there are likely to be more than one fora in the system there will also be opportunities for knowledge brokers that can inspect activities in different fora and help to cross-pollinate and assist knowledge transfer between them, when problems identified in one forum have a ready solution proposed by another forum. For the management of innovation it is important to construct social groups (contexts) where similar understandings and organisational behaviour emerges (i.e. Teams, Meetings, Operation workflows, formal and informal hierarchies etc.). By implementing new organisational methods as new roles supported by a semantic network, knowledge can be managed virtually in order to create an innovative environment. This cannot be solved by technology alone but requires active organisational effort by people to make the knowledge management solutions work to their best effect.
The CogNet solution will be comprised of two major elements, namely a personal knowledge organiser and its integration with a corporate cognitive network. The personal information and knowledge management tools will be implemented as a separate application resident on a PC and which will be integrated with a corporate intranet platform that supports the cognitive enterprise network paradigm. The implementation will address two aspects: Firstly new organisational methodologies are required, including newly developed roles. These roles will eventually be actively supported in the enterprise with dedicated personnel. Secondly the developed software will be implemented and extensive assistance will be provided to ensure easy user access and commitment.
The focus will initially be on the development of a platform, which supports ontology, based linking of concepts within the corporate knowledge domain, achieved by detailed analysis of the user domain and storage and access to the concept bank in a detailed user specific way. The pilot system will use an NT Server platform to support a readily available standard database package, such as Oracle. The eventual CogNet system will take the form of a personal organiser of information, suitable for storage on a notebook computer, in which will reside local copies of relevant public information, and user specific private information. The data stored on a notebook computer will be integrated with the central database when the user next connects including support for synchronising the data in the multiple storage areas, and keeping track of editorial control of documents in production. In order to facilitate easy integration and a common user interface between the off-line and on-line versions of the system, we will use a standard web browser interface for both the intranet side of the system. Public access will largely be denied, and also the extranet side where access to public and restricted parts of the database will be via the internet. The interface that we shall develop will be intuitive and graphical, and allow graphical point-and-click navigation around the system in terms of concepts and ideas.
5. Conclusions and Further Work
In conclusion the CogNet project is an experimental approach and a candidate solution for the rapidly growing importance of the need for intelligence and knowledge sharing, using a multi-disciplinary approach covering organisational behaviour and IT systems, making the human intelligence the design for a new knowledge management paradigm.
In this article we have examined the need and requirements for technologies for knowledge management that can be applied in an organisation at a corporate and at an individual level. We have also identified areas of related work and that will require further investigation. Finally we have proposed and outlined CogNet as a novel approach to knowledge management. We believe that this approach has considerable merit and potential and which needs to be developed further to fully exploit its potential. In particular the benefits of such an approach can be summed as follows:
References and Bibliography
[Ascari et al. 95] Ascari A, Rock M & Dutta S, "Reengineering and Organisational Change: Lessons from a Comparative Analysis of Company Experiences", European Management Journal, Vol. 13, Issue 1, pp. 1-30
[Bologna et al. 97] Bologna G & Gameiro Pais J, "ICARE: An Operational Knowledge Management System, Proceedings of the 5th Int. Symp. On the Management of Industrial and Corporate Knowledge (ISMICK?97), pp. 150-162, Compiègne, France, 1997
[Davenport et al. 98] Davenport T H & Prusak L, "Working Knowledge", Harvard Business School Press, 1998
[Dominique 98] Dominique J, "Tadzebao and WebOnto: Discussing, Browsing and Editing Ontologies on the Web", Proceedings of the 11th Knowledge Acquisition Workshop KAW?98, Banff, Alberta Canada, April 1998
[Eason et al. 96] Eason K & Olphert W, "Early Evaluation of the Organisational Implications of CSCW Systems", in Thomas P (Eds.) CSCW Requirements and Evaluation, pp. 75-89, Springer Verlag, London 1996
[Elliott 97] Elliott S.: "Arthur Andersen Maximises its Core Commodity Through Comprehensive Knowledge Management", Knowledge Management in Practice, American Productivity and Quality Center, No. 9, 1997
[Emerson 97] Emerson JC, "E&Y Sets the New Knowledge Standard Says Emerson", Emerson?s Professional Services Review Journal, March/April 1997
[Farquhar et al. 95] Farquhar A, Fikes R, Pratt W & Rice J, "Collaborative Ontology Construction for Information Integration", Research Report No. 63, Knowledge System Laboratory, Stanford University, Stanford USA, (ftp://kls.stanford.edu/pub/KSL-95-63.ps
[Grundstein 95] Grundstein M, "La capitalisation des Connaissances de L?entreprise, Système de production de connaissances", L?entreprise Apprenante et les Siences de la Complexité, Aix-en-Provence, Mai 1995
[Heijst et al. 96] van Heijst G, van der Spek R & Kruizinga E, "Organising Corporate Memories", Proceedings of the Knowledge Acquisition Workshop KAW?96, Banff, Alberta Canada, November 1996
[Huang 97] Huang KT, "Capitalising Collective Knowledge for Winning, Execution and Teamwork", (http://www.ibm.com/services/articles/intelcap.html)
[Kühn et al. 97] Kühn O & Abecker A, "Corporate Memories for Knowledge Management in Industrial Practice: Prospects and Challenges", Journal of Universal Computer Science, Vol. 3(8), pp. 929-954, August 1997
[Macaulay 96] Macaulay L, "Cooperation, Requirements Analysis and CSCW", in Thomas P (Eds.) CSCW Requirements and Evaluation, pp. 39-55, Springer Verlag, London 1996
[Nonaka et al. 95] Nonaka I & Takeuchi H, "The Knowledge Creating Company: How Japanese Companies Create the Dynamics of Innovation, Oxford University Press, 1995
[O?Leary 97] O?Leary D E, "The Internet, Intranets and the AI Renaissance, Computer, Vol. 30(1), pp. 71-78, January 1997
[Sowa 84] Sowa J F, "Conceptual Structures: Information Processing in Mind and Machine, Addison-Wesley, Reading, Massachusetts, 1984
[Sveiby 92] Sveiby KE, "The Knowhow Company: Strategy Formulation in Knowledge-Intensive Industries", International Review of Strategic Management, Wiley, 1992, pp. 167-186
[Sveiby 97] Sveiby KE, "The New Organisational Wealth: Managing and Measuring Knowledge-Based Assets", Berrett-Koehler Publ. San Francisco, 1997
al. 98] Tennison J. & Shadbolt N R, "APECKS:
A Tool to Support Living Ontologies, Proceedings of the 11th
Knowledge Acquisition Workshop KAW?98, Banff, Alberta Canada, April 1998