Reviewing the Literature to Guide AI Method Choice: Scoping Its Usefulness in the Context of Inventory Management Studies

Abstract

For practitioners, getting an idea of the AI methods that can be applicable to their organizational problems can be challenging due to the vastness of the field. Therefore, established development frameworks such as CRISP-DM, often recommend to find comparable use cases through a review of the literature. In this study, we assume the role of practitioners and emulate what we can or cannot learn from reviewing the academic literature about potential AI methods that are applicable to the exemplary field of inventory management. While we find that initial method choice can be informed by such a review (e.g., to narrow down the set of possible methods for popular use cases such as Demand Prediction), it also depends on how use cases are defined and distinguished from each other. Further, we find that there is often a lack of detail on why methods are adapted to a specific case context, which indicates the need for more standardization in the reporting of AI applications. Overall, a review of the academic literature can help practitioners to deal with the vastness of the AI field, though a more consolidated approach to the reporting of practical adaptations is needed in order for practitioners to be able to gauge whether a specific solution is also useful for them.