• Business Intelligence
• Collaboration
• Competitive Intelligence
• Communities of Practice
• CRM
• Culture
• E-learning
• Enterprise Content Management
• Enterprise Search
• Intranets & portals
• Taxonomies

• News
• Events

Feature

posted 3 Oct 2000 in Volume 4 Issue 2

Exploring the KM toolbox

With so many KM tools and portal products on the market, it is often difficult to fully understand the relationship between these applications and comprehensive knowledge management systems. By identifying the dominant fields of KM and artificial intelligence research, Eric Tsui develops a broader perspective of these technologies and discusses criteria for their evaluation.

People have disparate interpretations of the term knowledge management. While there is still no one universally accepted definition for KM, the general consensus is that to tackle knowledge management well, contributions from, among others, diversified areas spanning management, human resources, decision science, marketing, artificial intelligence and knowledge modelling are needed. By tracking relevant academic publications in these fields, three dominant streams of research into knowledge management can be identified.

The first stream focuses primarily on research into the theory of knowledge, the knowledge of the firm, organisational culture, measurement of intellectual capital and learning organisations. These research fields tackle the theoretical aspects of knowledge management, and develop models for valuing intellectual capital. Some researchers even challenge Nonaka and Takeuchi’s framework for the socialisation and externalisation of knowledge.

The second stream is represented by the work on corporate memories (aka organisational memory and organisational memory information systems) for enhanced decision making. A corporate memory embraces all forms of institutional knowledge, whether formally encoded within the current information systems, or tacit (informal) knowledge used by individuals in professional practice. This group has a strong focus on knowledge sharing and on practical applications of knowledge management in a corporate-wide perspective.

The third stream, with a strong contribution from computer scientists and artificial intelligence (AI) researchers in particular, tackles the areas of intelligent agents, ontologies (taxonomies), data mining, knowledge modelling, and computer-mediated collaborations.

Categorisation of KM tools

Especially so for the third stream, all the above research has spawned the development of tools for supporting various knowledge processes, for example capturing, encoding, organising, searching, distributing and measuring. The end product of using these tools (and more) to develop an application is a (technical) KM system. Generally speaking, a KM system is any computer system that integrates various knowledge processes in one or more organisations to solve specific business problems. In a wider sense, the objectives of developing and deploying a KM system are to:

• Capture, create and share knowledge assets
• Locate relevant information/knowledge
• Provide an environment for knowledge exchange
• Connect people with relevant interest and or skills
• Facilitate intelligent problem solving.

One can perceive from the above objectives that the field of AI, among other areas, has a strong influence on the evolution of KM tools (and hence applications). After nearly three decades of research, many of the relatively more mature techniques in AI have been packaged into commercial products. (One should also note that back in the 80s and 90s, large organisations invested considerably into the development of expert systems. These systems are usually problem solving in nature and are highly focused in a restricted domain. As a result, a great deal of decision-making knowledge has been acquired and represented in rules and or decision tables.) The following categorisation of KM tools, challenges and their respective trends are increasingly becoming evident in the market.

Intelligent search

By far, the majority of tools offer search capabilities. Search can range from simple keyword match, attribute-based input to context-sensitive search (that is, taking into consideration what the user’s interest, role type, and the very activities he or she is conducting just prior to issuing the search). Some tools also make extensive use of a word taxonomy or an ontology (which can either be manually created or automatically discovered with user guidance) to navigate the search space so that results are contained and are highly aligned to the user’s need(s). More sophisticated tools are emerging and these will, progressively, incorporate collaborative filtering techniques. That is, one can issue a goal statement (instead of keywords) to the search engine that translates the goal to a list of specific search probes. Collaborative filtering also allow search patterns, ideas and results to be shared (and reused) among a group of interested parties. One of the challenges in conducting searches is to properly synergise the result gathered from inside as well as external to an organisation.

Process modelling & mind mapping

These tools provide a visual environment for ideas to be captured and shared. Business processes can also be defined and modelled. Such tools greatly enhance the conceptualisation of procedural and factual knowledge. However, nearly all the tools in this category are stand-alone products. There are two key challenges to future tools in this category. Firstly, the ability to automate the conversion of the defined business processes to operable business objects for simulation allowing versatile questions to be posed. Secondly, set operations (for example, join, expand, contract, superimpose, conflict resolution, and so on) of concept maps need to be defined. For instance, there is no reason why email messages, which are mostly textual in nature, cannot be replaced in the future by concept or knowledge maps, so that knowledge workers can define and communicate visual information in synchronisation with a pre-defined corporate framework.

Case-based reasoning (CBR)

CBR is an AI technique that enables past cases (problems and solutions), with appropriate modifications, to be reused for unseen cases. Many CBR systems have been developed for the help desk, software development and CRM applications. With the increasing popularity of customer knowledge management, CBR tools will remain a dominant AI technique in the KM arena. On the research side, the relationship and synergy between CBR and organisational memory are actively being studied.

Data & text mining

This category of tools, which enable meaningful patterns and associations of data (words and phrases) to be identified from large databases, has been around for more than a decade. They form part of a KM solution as many developers and researchers consider data and text mining to be a type of ‘micro’ knowledge strategy (as opposed to knowledge programme management as a suite of ‘macro’ knowledge strategies) for an organisation. Data and text mining systems are being used extensively in business intelligence, direct marketing and customer relationship management applications. As most organisations only have a small group of data miners, it is doubtful that data & text mining tools, though they will undoubtedly remain a strong technical component, will be accessed via a enterprise wide corporate portal. In the near future, such tools will be gradually aligned with other tools to support key tasks in the above types of applications (for instance, from data gathering to data mining, encoding of business rules, capturing of decision making criteria, matching of customer profiles to product services, campaign management and the incorporation of feedback).

Web crawler

These are web-based tools that facilitate intelligent searching with extensive use of meta-data and indexing. Data is not limited to texts and numerals but is often in multimedia, for instance voice, graphics, video and so on. A common characteristic of web crawler tools is the ability to place ‘hooks’ on numerous locations on the web, monitor the content and activities on these pages and notify the user once there is change of content at those locations. Among other applications, such tools are especially suited for performing research on the web and gathering competitive intelligence.

Groupware

The two most dominant platforms are web-based intranets and Lotus Notes-based intranets. Detailed comparison of these two platforms has been widely undertaken and is not the focus of this article. While the web-based concept is extremely popular, generally more economical and compatible with nearly all of the tools on the market, Raven, the latest entrant from Lotus, has a very unique feature. By tracking user interest (and expertise), Raven assists in the location of relevant information, as well as the connection of certain employees in an organisation. Intelligent features like this and others will continue to be introduced by product vendors to gain differentiation and competitive advantage over their competitors.

Measurement & reporting

Some organisations tackle KM with a strong human resources focus and device criteria to measure the benefits of their KM programme. Tools are now available to measure, track and report on the value of intellectual capital (that is, non-financial assets) in an organisation. Most noticeably, these tools are based on the Balanced Scorecard method or the Intellectual Asset Monitor (ICM). Tools on tracking and reporting professional development, online self-paced learning, and performance reviews for employees are also available.

The above categorisation of tools is functional and techniques-based. Other categorisations also exist. For example, the Delphi Goup’s classification of KM tools extends to cover Enterprise Resource Planning (ERP), Electronic Document Management (EDM) and Information & Aggregration tools. IDC separates KM software into two groups: Infrastructure and access software. KM infrastructure software is the base or platform upon which KM solutions are deployed. Access software operates on the KM infrastructure to provide (groups of) individuals with access to (internal and external) knowledge repositories. Through fundamental research, Computer Sciences Corporation (CSC) has developed a KM Spectrum that maps KM systems into the following six categories:

• Transactional
• Analytical
• Asset management
• Process
• Innovation and creation
• Developmental (i.e. learning).

As a general observation of the KM tools on the market, they are generally both powerful and fast in the search for and dissemination of knowledge (i.e. documents and links). Some tools can cope with multi-modal information and most tools can handle files stored in various formats and in a range of platforms. Such tools are especially appealing to technical users and, on many occasions, have spearheaded the early adoption of a technical KM system in an organisation. However, this can also be a disadvantage for two reasons. Firstly, organisations should always define its KM strategies (or principles) before any tool is adopted (even though a particular tool is considered to be a good ‘technical fit’ to the problem in hand). Secondly, as in all acquisitions, organisations should always critically assess the ‘buy versus build’ proposition. Other potential shortfalls in KM tools are:

• They can be tedious to configure. Integration with legacy systems and operational databases has been reported as a key implementation issue.
• No one tool supports the full spectrum of KM processes.
• They can be poor in structuring knowledge and offer virtually nothing in understanding – most tools can only generate indices to facilitate subsequent retrieval. As an intelligent system, information or knowledge needs to be processed and understood. Natural language understanding techniques have yet to be incorporated into the commercial tools.
• They offer little support for knowledge reuse. Another consequence of the above is that knowledge cannot be re-organised and re-applied to a different problem. This capability, if available, is a key contributor to product and process innovation.

One of the key arguments for the last three shortfalls is that efforts should be directed towards developing advanced search engines rather than focus on trying to understand and classify knowledge in documents. Two justifications are that corporate knowledge management is different from personal knowledge management (where humans tend to classify things and store them with a view to locating them easily) and that one cannot predict the future usage of a piece of information, and hence it is pointless to think about creating an index for it without knowing how it will be used in the future. On the second point, artificial intelligence has long shown that indexing, representing and searching knowledge are intrinsically related. Furthermore, to be effective, an index has to be purposeful, general and predictive.

Irrespective of the ‘buy versus build’ decision, here are some useful criteria to apply when performing an overall evaluation of a KM tool:

• Power of the search engine (for example, intelligent search, collaborative filtering, efficiency)
• Flexibility and ease in browsing and availability of visualisation tools
• Automatic classification (using a taxonomy/ontology) of documents
• Collaboration and personalisation features
• Active collection and distribution of knowledge (i.e. knowledge pump)
• Repository for document management
• Ability to handle multiple file formats, media and technical platforms
• Cater for personal KM as well as corporate KM
• Capture operational statistics and provide analytical tools
• Extent and ease of access to external sources.

Portals

While KM tools have been around for years, it is the concept of the Enterprise Information Portal (EIP) that has gained immense popularity among business leaders in the last 12 months. Portals are seen as a one-stop entry point for staff, customers and partners to access and share information, to perform transaction(s), and to carry out specific work tasks. Portals are popular primarily because of the advance in e-business models (for example, cyber-stores, vertical integrators, net markets, volume procurement, and so on) and customer relationship management (CRM) (for instance, customer knowledge management, measurement of relationship capital and direct marketing). E-business has also impacted KM by gradually taking it from an internal focus (a corporate programme) to an external focus (addressing the flow of knowledge between/among suppliers and customers). As a result, many KM tool vendors have re-positioned their product offerings to align with the portal market (for example, commerce/trading, information, procurement, collaboration and learning portals). However, it is important to understand that the fundamental technologies underpinning the composition and hosting of portals remain predominantly the same.

Based on the above discussion on KM tools, and taking into consideration the forecasts by the Delphi Group and Gartner Group on portals, one can anticipate the following stages in the increased sophistication of future portals:

Stage 1

• Simple search mechanism and standard (static) page delivery
• Information dissemination point.

Stage 2

• Core content with pre-defined variations in page delivery
• Expanding set of interactive and informative facilities
• Centralised search facility for organisational content
• Engagement of common business processes.

Stage 3

• Advanced search facilities and seamless integration of search results (external and internal)
• Simple expertise locater
• Online knowledge communities
• Customers can initiate and check transactions
• Ability to perform transactions with suppliers and partners.

Stage 4

• Extensive and dynamic personalisation of content for individuals
• Active collection and distribution of knowledge
• Full integration with e-business systems
• Decision support and problem solving capabilities.

Some of the technical challenges to future portals are: Web-access to enterprise application systems; automatic categorisation of all corporate resources and assets; intelligent and multiple search strategies (for example, interactive and offline search, sharing of search cues and results); and, real-time user profiling and personalisation strategies.

Developing KM applications

Though not definitive, there are two major approaches to developing KM applications – process and product. A process-based approach (or ‘personalisation’) treats KM as a social communication process. The emphasis is on the deployment of an environment/

framework (whether technical or not) to foster better exchange of knowledge among staff and other participants. On the other hand, a product-based approach (or ‘codification’) focuses on the collection, storage and distribution of institutional knowledge for reuse. Ideally speaking, under the product-based approach, organisational knowledge is retained by a corporate memory and the memory is an active one (both in collection and distribution). A review of the reported KM applications has identified the following ‘tools of the trade’ adopted by organisations embracing KM:

• Cultural change
• Knowledge maps
• Ontologies/Taxonomies
• Corporate memory
• Expertise directory (or, people finder system)
• Online learning and performance tracking system
• Groupware (supporting online discussions and workflow, for example Lotus Notes, intranets)
• Authoring and content management tool
• Specialised (web-oriented) search and categorisation tools
• Data mining system
• Information repositories, for example documents databases, Electronic Document Management Systems.

Readers should note that the above tools are not mutually exclusive and organisations often adopt a combination of these to tackle a particular problem. With the above framework of KM tools and portals, one can develop a better appreciation of how these tools and capabilities benefit various KM applications. To conclude this article, five common KM applications are outlined with their respective KM supporting ‘components’ in figure 1. (Cultural issues and integration to legacy systems are undoubtedly key parts of any KM programme hence they are not repeated in the table. Readers should treat the information as a preliminary guide rather than a ‘blueprint’ to deploying a KM programme.)

It is hoped that this article provides the reader with a framework to better understand the types and origin of the KM tools on the market as they are still the very technologies underpinning the immensely popular concept of portals. If the latest product offerings by the tool vendors are a guide, then many vendors are now offering the above tools, and more, to develop portals to cater for major activities in KM, e-business and CRM applications. KM

Eric Tsui is Chief Research Officer, Asia Pacific of Computer Sciences Corporation and an honorary associate of the University of Technology, Sydney. He can be contacted at: etsui2@csc.com.au or etsui@it.uts.edu.au