Design intent
SAP remains the system of record. The proposed platform ingests operational data and creates historical/features datasets for forecasts, monitoring, explanations and recommendations.
MVP scope
The repository documents demand forecasting, supplier lead-time prediction, shortage prediction, exception prioritisation and read-only/shadow operation. A planner control tower is described as the intended interface.
Governance boundary
AI is not intended to modify SAP directly. Business rules and human approval are prerequisites for any future controlled write-back. Stock optimisation, capacity risk and scenarios are documented as later extensions.
Technology status
Python/FastAPI, PostgreSQL, Redis, conventional ML libraries, MLflow, optimisation tooling, React/Next.js and containers are described as technology directions. They are not independently verified installed components of a running production platform.
Source: https://github.com/tabshoraailab/mrp-intelligence-platform ↗