ChainSink

The MVP Product Document for XAI in Supply Chain Optimization with IBM Granite presents a lightweight yet impactful prototype designed to enhance transparency and decision-making in supply chain management. This demonstrates how explainable AI (XAI) can improve forecasting, inventory management, and risk assessment in the textile and retail industry. The solution focuses on four core features: demand forecasting, which predicts future sales while providing clear justifications (e.g., seasonal trends, historical sales patterns); inventory optimization, which recommends ideal stock levels based on lead times, demand fluctuations, and supplier reliability; a risk alert system, which proactively flags supply chain disruptions with real-time root-cause analysis; and a scenario simulator, allowing users to tweak key variables (e.g., demand increase, supplier delays) to assess potential impacts dynamically. The technical stack includes IBM Granite for AI explainability, Python and Pandas for backend data processing, Streamlit for a simple and interactive front-end, and Plotly/Matplotlib for advanced data visualizations. Users can upload CSV/Excel datasets (or use sample data) to generate real-time insights into sales trends, stock levels, supplier risks, and demand forecasts. Designed with a minimalist, data-driven UI supporting dark/light modes, this prototype is tailored for quick decision-making in supply chain networks. Future development plans include integration with ERP systems (SAP, Oracle), multi-user authentication, and expanded AI-driven logistics optimization, ensuring scalability and deeper industry impact.