C/02: “Tomorrow's Spreadsheets : Conceptual Graphs as the Knowledge Based Decision Support Tool for the Management Accountant”

This paper presents conceptual graphs as the means by which management accountants are able to evaluate qualitative business problems. Management accountants can, using conceptual graphs, structure concepts and relationships beyond that of numeric and data-based decision support tools such as the spreadsheet, which triumphed because users were conversant with them manually. The same success is possible with conceptual graphs. A small example shows how conceptual graphs can benefit the management accountant. The paper assumes the reader is unfamiliar with management accounting. However, the paper presumes a fundamental understanding of conceptual graphs theory. Introduction Management accountants strive continually to provide timely, accurate and meaningful performance gauges to business managers. Yet, although current decision support tools such as spreadsheets are flexible, their conceptual roots lie in numeric and data-based relationships. Many important business problems demand solutions based on subjective judgement. Many writers, like Graff (1989), acknowledge the fact that these qualitative evaluations still have to be processed and explained manually. A large body of current evidence, catalogued by Mahahon (1990), suggests another approach which should be knowledge-based. This is demonstrated by the emergence of accounting expert systems. Despite this, any successes to date have been restricted to solving a specific range of easily definable problems. Much of the literature, typified by Keravnou and Washbrook (1989), discuss this limitation in terms of overcoming the technical inflexibilities within the various expert system formalisms. Although this is a worthwhile pursuit, there is a better answer to this issue than seeking improved knowledge representation formalisms alone. Gill (1988) argues convincingly for a user-centred approach in expert system research and development. Technically-centred expert systems attempt to