In the author’s own words, this book “gives an introduction to knowledge-based systems, and how to develop these systems in a wide range of business situations.” It is split into three parts. Part 1 is the expected introduction to expert systems; it covers knowledge representation, inference, knowledge acquisition, languages, and shells. Some of the material is very shallow; certainty factors, for instance, are covered in one paragraph. On the plus side, the author goes into some depth on how to select an appropriate application, analyze it, and develop a prototype. Three chapters provide an introduction to six widely available shells: M.1, EXSYS, LEVEL FIVE, 1st-CLASS, VP-EXPERT, and GURU. Discussion of languages, such as Prolog, and environments, such as KEE, is again very shallow.

Part 2, is made up of 15 example applications, each occupying a single chapter. Most applications are financial (e.g., loan approval, investment planning, external auditing) or strategic (e.g., new venture planning, marketing strategy). Each chapter starts with an overview of the problem area, discusses the selection and analysis of an application, and concludes with all or some coding from a prototype in M.1. Teaching expert systems by exposure to numerous worked examples is appealing, and certainly better than the overworked method of discussing the ‘classic’ applications. Unfortunately, the approach taken here is flawed. First, none of the examples are truly problem-driven--they are all concerned with finding a role for the technology irrespective of any potential use or user. Second, all knowledge is presented as used in the example, and the reader gets no feel for knowledge acquisition. Third, there is no discussion of validation nor any attempt to evaluate how useful the resulting system might be. Fourth, each example is tackled and presented in a very similar manner, leaving the reader with the impression that expert systems development differs little between applications.

Part 3 contains six previously published papers by various authors, covering available shells, selection of a shell, knowledge acquisition and elicitation, and two applications. The rationale for Part 3 is not clear: all of the papers are commonly available elsewhere.

The market for this book is no doubt introductory courses on expert systems in business schools, particularly for information systems specialists. I teach such a course but will not use this book. The shortcomings of the introductory material are too numerous, and the example applications convey an unrealistic approach to development. The text is also far too large, consisting of 640 pages. Removing Part 3, dropping most of the example applications, and rewriting much of the introduction would result in a better book of well under 300 pages.