This dissertation describes the development of a knowledge-based system, dubbed the "Intelligent Facilitator's Assistant" (IFA), which is capable of assisting a facilitator with selected aspects of meetings employing tool-oriented electronic meeting support (EMS) systems. Issues in the separation of site-specific and domain-specific knowledge are discussed, as are alternatives for modeling facilitator behavior. Software activities carried out as part of the research, which included the construction of a general-purpose microcomputer-based expert system development environment, are also covered. The final products of the research are conceptual models of product-oriented work groups and expert facilitator behavior in pre-session planning; these are realized in a taxonomy of products for EMS sessions and in a prototype knowledge base for agenda construction. Significant conclusions of the research are: (1) an intelligent assistant must be based on deep knowledge of the tool environment; (2) such an assistant must accommodate multiple models of both facilitator and user behavior; (3) the advent of intelligent assistants has major implications for the design of software for this environment, specifically in providing for control and visibility of tool activity by a process rather than a human operator.
展开▼