Assessing impacts of information systems in general and knowledge based systems (KBSs) in particular is still largely based on anecdotes and folklore, leading to support mainly in the shape of checklists. Though there is nothing wrong with this pragmatic approach, there is a need to put the issue on a firmer, scientific, foundation. This paper describes research aiming at this goal. It defines a number of concepts and a modeling formalism which is used to develop a (limited) theory about the relation between properties of KBSs and properties of the production process in which they are used, and changes in this process and the changes in costs incurred for this process. The hypotheses derived from the theory are tested on a set of 13 KBSs which have been deployed during the last decade. The main conclusions of the paper are that the general approach followed clarifies problems in impact assessment and that to a certain extent the properties of activities that make up the production process in which the KBS is used determine the change in cost and changes in the way the process is organized. Factors related to properties of the KBS seem less predictive in this respect.
- Knowledge based systems