An expert system approach to personal financial planning in Australia

The objectives of this research were to investigate issues involved in the development of a financial planning expert system for the Australian environment and to test the system's effectiveness as a decision support tool. A prototype expert system was developed and tested outside the development environment. Findings indicate that it is possible to develop a simple rule-based system which gives advice useful to practitioners in an area of personal financial planning, namely, advice on superannuation. Findings also indicated that expert systems help to overcome several kinds of representational failures in the human analysis process including: inconsistency, inaccuracy, individual differences, and inefficiency. Those aspects of expert systems that directly support a decision maker in reaching an effective decision were identified. These aspects were: intelligent calculation capabilities, explanation as to why certain information was required by the system, and explanation as to how conclusions were reached by thesystem.From a general perspective the methodology used in this study offers the foundations needed to attest the value of expert systems as effective decision support tools. Future studies should incorporate a longitudinal design to more accurately assess cumulative effects of an expert systemin decision making. This study was of an exploratory nature and the major limitation of the study was the small number of test cases studied. However the study is of significance as there are very few reported studies evaluating the effectiveness of expert systems in decision making. Thesuccessful verification of the system developed demonstrates its feasibility. The main messages to expert system developers and researchers from this study are that expert systems will flourish more by focusing on 'how' theycan be improved as decision tools and that there is a clear need for more attention to theoretical grounding than is generally assumed or practiced in expert system development and research. Expert system research should also be integrated with Decision Support System [DSS] design literature since, as was found in this project, understanding which DSS are effective and what makes them effective is of great assistance in improving the designof expert systems.