Predictive analytics and machine learning (ML) are reshaping supply chain management by converting fragmented operational data into forward-looking decision intelligence. This paper examines how demand and lead-time forecasting, anomaly detection, supplier risk scoring, and prescriptive optimisation improve end-to-end visibility across multi-tier networks, enabling earlier disruption sensing and faster response. By integrating data from ERP and procure-to-pay systems, transport and warehouse platforms, IoT sensors, and external signals such as weather, geopolitical risk, and commodity indices, analytics models enhance situational awareness from suppliers to customers. The study explains how probabilistic forecasts support agility through dynamic inventory positioning, capacity rebalancing, and route reconfiguration, while optimisation models translate predictions into actionable policies that protect service levels under uncertainty. Profitability gains are analysed through reductions in expediting, stock-outs, obsolescence, and safety-stock inflation, alongside improved working-capital efficiency and margin protection via better cost-to-serve decisions. The paper also outlines how analytics-enabled executive decision processes improve accountability by linking model outputs to operational KPIs, scenario stress tests, and governance routines that monitor bias, drift, and data quality. The paper concludes with governance and KPI guidance for adoption.