Abstract

Organizational decisions have become more data-driven and collaborative, with the increasing utilization of artificial intelligence, machine learning, and analytics to support decision making. While humans and machines are each bounded in their own rationalities, their collaboration has enabled a new, collaborative rationality by augmenting the intelligence and capabilities of each. New research is required to explore the degree and mode of human-machine collaboration, with the aim of enhancing collaborative rationality, and its effect on decision making. Furthermore, the resulting decisions must be evaluated to enable learning, rationalization, and sensemaking from the decision outcomes. However, data-driven decisions are complex in nature, and current theoretical developments fall short in accommodating for their multi-faceted nature and changing context, and there is lack of theoretical support on how, when, and why to evaluate such decisions. Accordingly, we follow a design science research methodology to develop and evaluate a model for data-driven decision evaluation. This model depicts the relationship and role of the multiple elements involved in data-driven decision making, and provides a feedback loop which inputs the results of evaluating decision outcomes back into the process/system, thus enabling learning from the past through experience, and ultimately enhancing collaborative rationality and decision making.