The realities of the past that are embedded in the repository and shared by later groups constitute the collective memory. This paper outlines the specification of a collective decision-making system that utilizes the work of prior groups stored in the collective memory. We describe an approach that uses group memory based reasoning. In an effort to enhance utility, our group memory system is tested in a bank loan analysis domain, a semi-structured decision-making environment involving multiple attributes. The main goals are to reduce the time required to come to a decision, particularly, in a critical situation, to compensate for lack of experience of young bank consultant, and to distribute available experience to different sites.