One of the reasons knowledge systems are justified is that the functionality being computerized is complex. But one of the reasons that applications become complex is that their requirements are constantly changing (e.g., underwriting policies, tax laws, pricing policies). Although KBSs may in some cases be easier to change than conventional code, the cost of such changes can ruin the economics of an application. Furthermore, due to high functionality, the number of users of any KBS application seem to be small, and any costs must be retrieved from the benefits to this small number of users.
KBSs are not, in general, a complete solution, but must be combined with other technologies to solve almost any real problem. The KBS approach is more of a technique and a skill than a solution. Thus when we try to assess the value of a KBS, we are often looking at too small a grain size to speak of the cost-benefit of an application. The NKK representatives made this point very explicitly. When choosing techniques to solve a problem, such as choosing between C or LISP, or between linear programming and a KBS, one cannot normally attribute the costs or benefits to the technology employed. These attributes belong to the application (which probably uses many techniques).
There are business opportunities in selling shells, selling very broadly applicable and relatively static applications, and selling services for the development of high-valued solutions. The overall market, however, does not appear to be very large when measured against databases, spreadsheets, etc. This situation is the same in Japan as it is in the United States.
Knowledge-based systems at their present technology level are more likely to be a core competence -- a technique applied when appropriate, commingled with many of the more conventional data processing techniques, and applied to the internal processes of a business or as an embedded component of a product. Clearly, we need a next generation of technology to fully realize the enormous potential of AI; then we will probably see an exponential type of business growth. However, given the investment practices of the United States vs. Japan, the latter (both the JCMs and the companies using the technology) would appear to be in a better position to realize the benefits of new research results.