More and more data is available to improve supply chain decision making but it needs to be carefully applied considering human limitations. Against this background, I will present two studies that focus on the role of human judgment in supply chain management decision making, first exploring the influence of planners’ adjustments to AI-generated demand forecasts and second examining the effectiveness of human decision-making in inventory management subject to inaccurate data. Study 1 investigates the role of human judgment in demand forecasting. We analyze planners’ adjustments to AI-generated forecasts using a dataset containing 30 million SKU-store-day level forecasts and associated variables. We employ random forest and decision tree approaches to understand the drivers and quality of adjustments. Our findings suggest product characteristics such as price, freshness, and discounts are important factors in adjustments. Large positive adjustments are frequent but often inaccurate, while large negative adjustments are accurate but less common which indicates behavioral biases. In Study 2, we focus on decisions made under the inaccurate inventory data due to shrinkage and loss. We explore the trade-off between cleaning inventory data centrally and allowing decision makers to adjust ordering decisions based on their judgment. In light of human biases in decision making, we present a set of hypotheses on the cleaning-adjustment trade-off and test them in a laboratory setting. The study raises questions about the effectiveness of normative models in determining whether to clean data centrally or rely on decision makers’ judgments, providing insights into optimizing human knowledge utilization in supply chain management.