A knowledge-based electronic messaging system: Framework, design, prototype development, and validation.

Although electronic messaging systems (EMS) are an attractive business communication medium several studies on the usage and impact of EMS have shown that despite the benefits, they have been generally used for routine and informal communication activities. Theoretically, EMS have yet to find their niche in organizational communications. Technically, EMS designs are not flexible to support communication activities of managers, are not maintainable to permit easy integration with other office applications and access to information from data/knowledge bases, and are not easily extendible beyond the scope of their initial design. Behaviorally, end users are not directly involved in the development of EMS. This dissertation attempts to bridge the transition of EMS technology from message processing systems to communication support systems. First, the dissertation provides an analysis for a knowledge-based messaging system (KMS) through a framework. The framework provides a theoretical basis to link management theory to EMS technology. It suggests that the communication needs of the managers vary depending on the activity level, implying related variations in EMS functionality. Second, the dissertation provides a design for the KMS through an architecture which incorporates the design and implementation issues such as, flexibility, maintainability, and extendibility. The superimposition of the KMS on an existing EMS provides flexibility, the loose coupling between the KMS-interface components and the KMS-functions increases its maintainability, and the strong functional decomposition and cohesion enhances the extendibility of the system beyond the scope of its initial design. Finally, the dissertation provides a implementation through the development of a prototype KMS which involves users into the design process through a validation study conducted at University of Arizona. The prototype used GDSS tools in eliciting message attributes for the personal knowledge base. This method proved effective in reducing the bottleneck observed in the acquisition of knowledge from multiple experts, simultaneously. Similarly, the combination of observation with interviews proved effective in eliciting the organizational knowledge base. The validation method measured the system's accuracy (which was very accurate) in prioritizing messages for the users