Decision Support as Knowledge Creation: A Business Intelligence Design Theory

Introduction

The purpose of decision support systems (DSS) is to improve the quality of decisions (Holsapple & Whinston, 1996; Keen & Scott Morton, 1978; Turban & Aronson, 2005). Since decisions are ultimately based on the decision maker’s mental model(s) (Argyris & Schön, 1996; Weick, 1995), improving decisions is achieved by updating and enhancing the decision maker’s mental model (Johnson-Laird, 1983; Kaplan & Kaplan, 1989; Qin & Simon, 1995).

Mental models are enhanced through knowledge creation (Nonaka & Takeuchi, 1995; Simon & Longley, 1981); i.e., through the combination and assimilation of relevant explicit information (e.g., trade journals, group discussions, new facts, written procedures), and by the creation of new knowledge through tacit mental analysis (e.g., evaluating/comparing mental models, generating insight(s), integrating ideas, developing causal relationships) (Alavi, 1994; Nonaka, 1994; Nonaka & Takeuchi, 1995; Senge et al., 1994).

Given that business intelligence (BI) technologies provide historical, current, and predictive views of business operations to enhance the comprehension/understanding of fact-based interrelationships, we suggest that BI is an appropriate enabling technology for knowledge creation. Specifically, we propose and develop a business intelligence design theory (BIDT) for DSS as knowledge creation. This BIDT provides a prescriptive theory that indicates how a design process can be carried out in a way that is both feasible and effective. Our BIDT addresses both the product of the design (i.e., the class of systems being designed -- DSS) and the process of the design (i.e., the selection and application of the appropriate BI technologies to knowledge creation in a DSS). While BI is normally focused on large internal (or external) data sources (e.g., data mining in fraud detection, competitive intelligence), our BIDT focuses BI internally on the decision maker, and on the discovery and enhancement of his/her mental model(s).

This article’s primary contribution to the literature is twofold. First, it proposes a unique BIDT for DSS as knowledge creation, based on Nonaka’s (1994) knowledge spiral, including separate kernel theories for each of his four types of knowledge creation. Kernel theories are theories from the natural or social sciences that govern the goals and construction of the system; i.e., theories that address the requirements for improved mental models, knowledge creation, and decision making using a DSS. These kernel theories and the associated process design methods are drawn from, and integrate components of, five distinct literatures: cognitive science, the theory of learning, artificial intelligence, knowledge management, and decision theory.

The second contribution of the article is the identification and application of several instance-based research systems, and the associated BI technologies, to the design methods of the BIDT. The integration of these instance-based systems demonstrates one way of building a DSS as knowledge creation.

The article is organized as follows. In the next section we propose the BIDT product design for decision support as knowledge creation. In the following section, we propose process design kernel theories for each of the four types of knowledge creation in Nonaka’s knowledge spiral, as well as testable design process hypotheses. Then, we provide an example of an instance-based DSS as knowledge creation that satisfies both the product design hypothesis and the process design hypotheses of our BIDT. And finally, we discuss several potential applications and research directions of DSS as knowledge creation.