Szendrey, Jaszmina (2016) Facilitated Expert System Modelling to Support Organisations in Making Strategic Decisions, TAF, Glasgow, United Kingdom
Strategic decision making typically entails dealing with ill-structured, knowledge-rich, non-quantitative and ambiguous problems which require simultaneous handling of a large amount of knowledge from different areas of expertise. Such decision making usually involves widely disparate interest groups with multiple, possibly conflicting goals and priorities. In search for suitable tools and methods to address the challenges of knowledge-rich problem domains some operational researchers see potential in expert systems. However, generally, the process of expert system development can be characterized as ad hoc and fails to address the challenges of facilitation. This thesis addresses these shortcomings by developing the process of Facilitated Expert System MOdelling (FESMO) using a particular knowledge-based expert system shell, called Doctus. The development of the FESMO process was grounded in both the principles reported in the literature as well as the requirements of Doctus, and was gradually refined through the findings of the empirical investigation conducted in nine European companies from the ICT field and brainstorming sessions held with the experts of Doctus modelling. Three complementary strategic decisions were chosen from the family of sourcing decisions (shared services, outsourcing and buying decisions) for modelling. Additionally, the decision models developed for particular decision situations of the case companies were subsequently integrated into a generic Business Process Sourcing Decision Model (BPS DM). In order to gain a better understanding of the applicability of the developed FESMO process and to develop a generic BPS DM, I designed a collective-case study research process with elements of ethnography and action research. To improve the reliability of the research findings I designed this research process as a multi-method qualitative research with data triangulation.