Expert Systems: An Integral Part of Knowledge Management

Feature

posted 1 Jan 1998 in Volume 1 Issue 3

Expert Systems: An Integral Part of Knowledge Management

Dr. Jay Liebowitz, Professor, Department of Management Science, George Washington University explores the under utilisation of expert systems in the knowledge management structure.

Expert systems are a powerful technology, but they haven't caught on as widely and quickly as people originally thought. Even though they are being used in most countries throughout the world (see The 4th World Congress on Expert Systems web site at http://www-cia.mty.itesm.mx/wces98; March 16-20, 1998 in Mexico City), albeit at different levels of maturity and application, expert systems have been one of the best kept secrets. With the rapid emergence of knowledge management, expert systems can find a secure home in which to flourish and become an integral and integrative element of knowledge management (see Liebowitz, J. and L. Wilcox, eds., Knowledge Management and Its Integrative Elements, CRC Press, Boca Raton, FL, 1997).

Knowledge management generally consists of four functions: securing, creating, retrieving/combining, and distributing knowledge. Much of knowledge management is not new. Much of its roots can be found in the expert systems and artificial intelligence fields. For example, the knowledge acquisition phase of expert systems can be easily applied to the capturing and securing of knowledge. Developing knowledge repositories for knowledge management activities can be easily traced to knowledge representation and knowledge encoding methodologies and techniques in the expert systems field. The indexing of knowledge can be traced to even case retrieval, similarity, and adaptation methods applied in the case-based reasoning area of the expert systems field. Thus, much of the underpinnings of knowledge management is derived from earlier work in the expert systems and artificial intelligence field.

Even more important than acknowledging part of the roots of knowledge management coming from the expert systems field is the realization and understanding that expert systems should be an integral part of a knowledge management system. Capturing expertise and putting it on-line in terms of on-line pools of expertise or web-based interactive knowledge centers is critical to the potential success of knowledge management. For example, in the November 1997 issue of the ABA Journal, David Vandagriff (Director of Technology Alliances for Lexis/Nexis) says that, 'We see it [the Lexis/Nexis Exchange] becoming an on-line legal community that will include expert systems with Web-based engines using artificial intelligence. For instance, there will be a federal court venue expert. After a user responds electronically to a few questions, an answer will be given concerning proper venue (p.84).' The US Department of Labor (www.dol.gov) already has been developing web-based expert systems (e.g., determining Veteran's benefits) as part of its knowledge compliance/knowledge management systems. Other companies and organizations are following suit to allow expert systems to play an important role in their knowledge management system, but still many others are lagging behind.

Knowledge management involves understanding how the enterprise works. In an early article titled, 'Knowledge Management: A Fit With Expert Tools,' that appeared in the November 1990 issue of Software Management, William Stapko recognized early on that expert systems and other AI tools can greatly impact knowledge management activities. He states, 'Managing knowledge is a high-level corporate concern. Management wants to know how to run and manage a business using rules and guidelines to reference everything from marketing to manufacturing. Expert systems provide the ability to insulate the business knowledge from the technical knowledge (p. 63).'

Expert systems and other artificial intelligence technologies have been maturing over the years. According to Sara Hedberg in her article, 'Where's AI Hiding?' (AI Expert, April 1995), she indicates that AI may be hiding in many little-known places, but it is alive and kicking. Mrs. Fields Cookies, Disney Store, IRS, Microsoft Word, the White House, Xerox, Compaq, and many other organizations have used expert systems to assist them in their activities.

So the question remains, why don't more knowledge management officers recognize the importance and need for expert systems within their knowledge management structure? In speaking with the Director of Knowledge Management at a well known organization in Washington, D.C., he failed to see the significance of using expert systems and their underlying methodologies for his company-wide knowledge management effort. He said that expert systems didn't seem to work in his organization when they were introduced years ago. Ever since then, expert systems have had a bad taste and we prefer to not use them.

I explained to this Director that he is missing the boat and expert systems technology (when applied to the appropriate problems and when expectations are not oversold) has matured to where it is a critical technology and business solution for many organizations. He did not seem convinced, however, and I'm afraid that he, and many other Chief Knowledge Officers or those with similar titles, may be underestimating the worth of expert systems usage within the enterprise knowledge management structure.

In Michael Dertouzos' 1997 book titled What Will Be: How the New World of Information Will Change Our Lives, he mentions that another probable organizational development is the evolution of 'expert centers' staffed by groups of related experts capable of high-quality, high-speed work at very competitive prices. Expert systems have a great role to play here. In fact, using expert systems as these 'expert centers' for on-line expertise and help should hopefully be part of the knowledge management organizational system, if knowledge managers realize a true potential of expert systems.

In Liebowitz and Beckman's forthcoming book Knowledge Organizations: What Every Manager Should Know (CRC Press, Boca Raton, FL, in press), they describe an eight step process for knowledge management. The stages are:

Stage 1: IDENTIFY

Determine core competencies, sourcing strategy, and knowledge domains

Stage 2: CAPTURE

Formalize existing knowledge

Stage 3: SELECT

Assess knowledge relevance, value, and accuracy; resolve conflicting knowledge

Stage 4: STORE

Represent corporate memory in the knowledge repository with various knowledge schema

Stage 5: SHARE

Distribute knowledge automatically to users based on interest and work; collaborate on knowledge work through virtual teams

Stage 6: APPLY

Retrieve and use knowledge in making decisions, solving problems, automating or supporting work, job aids, and training

Stage 7: CREATE

Discover new knowledge through research, experimenting, and creative thinking

Stage 8: SELL

Develop and market new knowledge-based products and services.

Within this framework, expert systems could be used in the Store, Apply, and Sell stages. According to Tom Beckman of the Internal Revenue Service, the field of AI is instrumental in many of these innovations. Business value-added comes from identifying and applying expert systems in situations where expertise and knowledge are required to solve problems. Knowledge engineers elicit expertise from domain experts and organize and structure it in ways that can be stored and applied in active forms to structure, guide, perform, and manage tasks; solve problems; and make decisions. The AI disciplines, and especially expert systems, can support the knowledge management process.

Expert systems can also be used as the integrative element linking various knowledge sources. They can serve as the integrative mechanism for solving interdisciplinary problems.

Expert and knowledge-based systems provide the framework for handling the exchange and integration of knowledge from various sources. They allow knowledge bases to be created for ultimate sharing and analysis. They are an ideal technology for capturing, preserving, and documenting knowledge, especially in today's environment where organizations are reengineering, downsizing, and losing senior managers due to early retirement packages. Expert systems can be very useful for building the institutional memory of the organization before this intellectual capital is lost.

According to Professor Dan O'Leary of USC in his article, 'The Internet, Intranets, and the AI Renaissance' (IEEE Computer, January 1997), shared knowledge is at the core of organizational or group memory and is essential to the preservation of expertise or process knowledge. ARPA's Intelligent Information Services project has moved to support virtual groups with a number of emerging technologies, including: institutional memory tools that help organizations capture expertise, including process knowledge and access to expert consultants; tools to support multiuser/multiauthor hypermedia Web development so groups can build their own Web sites; and self-organizing knowledge repositories that adapt to community needs with use. As an offshoot of expert systems, knowledge-sharing agents are starting to emerge which could facilitate the knowledge management process.

In the July 1997 special issue on 'Knowledge Management' in the Expert Systems With Applications International Journal (Elsevier, Vol. 13, No. 1), knowledge-based and expert systems are noted for having the potential of playing a major role in the knowledge management era. Karl Wiig of the Knowledge Research Institute in Arlington, Texas points out that historically, the major impact of knowledge-based system applications in support of knowledge management has been to deliver knowledge to the point-of-action' where the most accurate information on the situation normally is present, analysis is performed, decisions are made, and the opportunity to serve the business in a timely manner is best. However, at the present time, an increasing number of knowledge-based system applications can take on other roles such as to build and organize knowledge, to support education, and many other purposes.

In Holsapple and Joshi's paper on 'Knowledge Management: A Three-Fold Framework' performed for the Kentucky Initiative for Knowledge Management, it is quite evident that expert systems can play a role in their framework of a knowledge resources component, a knowledge management activities component, and a knowledge management influences component. Certainly, expert systems could aid in the knowledge management activities component by representing and processing knowledge.

So what is the message here? The key message is that expert and knowledge-based systems should be recognized by knowledge managers as playing a fundamental role in the development of the organization's knowledge management system. Knowledge management is not new - knowledge managers and knowledge analysts should lift up the outer lining of the knowledge management coat and examine what's underneath. They will quickly find that many of the methodologies, techniques, concepts, and tools from the expert systems and AI field can be appropriately applied to knowledge management. Knowledge management is the new term in vogue now for repackaging many of these ideas which developed from the AI/information technology, organizational behavior, and human resource management disciplines. As knowledge management tells us, let's learn from the past and these other disciplines so that we don't reinvent the wheel. A great way of doing this is to apply expert systems technology to capture these lessons and further apply this knowledge in a pro-active manner. Use the power of expert systems technology to the knowledge management field and there will be a greater likelihood for success of the knowledge management era.

Professor Jay Liebowitz is a Professor at the Department of Management Science, George Washington University. He can be reached at:

jayl@gwis2.crc.gwu.edu.

His recent book titled 'Knowledge Management and Its Integrative Elements' (J. Liebowitz and L. Wilcox, eds.) was published by CRC Press (Boca Raton, FL) in June 1997. His forthcoming book entitled 'Knowledge Organizations: What Every Manager Should Know,' (J. Liebowitz and T. Beckman) will be published by CRC Press in May 1998.