Abstract
Interpretability in Business Intelligence systems (BIS) refers to the ability of end users to make sense of the information and insights generated by BIS. The increase in large datasets, complex analytics, and the emergence of Artificial Intelligence (AI) make BIS interpretability increasingly important for decision-making within organisations, yet there is a lack of theoretical guidance on evaluating interpretability in the BIS context. Therefore, this paper aims to identify evidence-based constructs for evaluating BIS interpretability. The research design includes a critical review of the literature to identify BIS interpretability constructs and related criteria, followed by a survey of BI experts and users to explore the current perspective on interpretability in BIS, the challenges faced in the retail sector of New Zealand, and the relevance of the literature-based constructs to industry experience. The theoretical contribution is the literature-based, empirically validated constructs, which can be operationalised through questionnaires to assess the effectiveness of BIS interpretability in practice.
Recommended Citation
Wyk, Quintus van; Biljon, Judy van; and Poel, Etienne Van Der, "Empirically validated constructs for evaluating the
Interpretability of Business Intelligence Systems" (2025). ACIS 2025 Proceedings. 52.
https://aisel.aisnet.org/acis2025/52