Delayed supplier deliveries or supply chain disruptions can cause production problems and lead to financial loss. If supply chain leaders can predict these disruptions rather than react to them (reactive firefighting), it will help avoid losses. 88% of organizations are experiencing these delays, and 60% of them have significant revenue loss (>15%) as a result. Traditional ERP systems provide historical data that can support retrospective reporting, but this may not be useful enough to predict supply chain delays. This paper proposes a proactive AI/MLdriven predictive model that can analyse historical Purchase Order (PO) data to identify high-risk orders before they are delayed. This early warning of supply chain delays helps reduce stockouts, improves on-time delivery rates, and provides data-driven decision support for procurement teams. In this paper, AI techniques such as machine learning and predictive analysis are used to analyse historical PO data, train and tune predictive models, and facilitate deployment. It portrays a new architecture that embodies AI capabilities within existing ERP frameworks, with a specific emphasis on modular adaptability and scalability. This research concludes with a discussion of challenges and limitations, and further directions toward the smooth adoption of AI in enterprise environments.
Sampath Kumar Mucherla (Mon,) studied this question.