Given that user involvement is important in system design, this study addresses the as yet unexplored question of how a user should be involved in the system design process. Two radically different processes of interaction between a systems designer and a manager were compared in an information system design exercise. In one interaction process, the designer conducted a traditional interview of the manager. He asked questions, analyzed data and made suggestions. In the alternative interaction process, there was an initial sharing of information and mutual suggestions, followed by a critique of each other's suggestions. The study found that for this ill-structured problem: The alternative interaction process produced higher quality designs with important implementation advantages. The two interaction processes produced designs which used different types of organization control strategies. These results may be due to a problem finding contingency—different processes of interaction may help to define different problems, and thereby produce different, but equally rational, solutions. For the management of organization design, the process of interacting with users has important implications for the quality of the resulting design, the type of organization control strategy employed, and the subsequent implementation of the system.