An experiment to assess the performance of a redesign knowledge system

Abstract

Today, a second generation of computer-based reengineering tools employs knowledge systems technology to automate and support key intellectual activities required for effective process redesign. But a central question remains as to the effectiveness of redesign through such knowledge systems. The research described in this paper is focused on testing the effectiveness of knowledge-based, process-redesign systems. We employ one such system, called "KOPeR-lite," as a platform for experimentation to assess the relative efficacy of redesigns generated by computer versus those developed by people. In this sense, we conduct a modified Turing Test to compare redesign performance of reengineering analysts with that of the knowledge system. KOPeR-lite performs comparatively well in certain respects, but human subjects outperform the machine in others. The results provide evidence to support claims of redesign efficacy through knowledge systems, and they offer insight into the relative strengths and weak nesses of people and software applications in the reengineering domain. This study further opens up new lines of research and highlights implications for process redesign and practice, including issues associated with leading adoption of knowledge system technology and extension of redesign automation systems such as KOPeR-lite.