Updating planning parameters is a practical first step for supply-chain AI

Gregory Pitstick argues that one of the fastest ways to generate value from supply-chain AI is to improve the parameters already driving ERP and planning systems, rather than beginning with a large platform replacement.

The article identifies five priorities: supplier lead times, safety stock, reorder points, min-max levels and lot-sizing rules. These parameters frequently drift as supplier performance, demand patterns, product mix and service expectations change, while the underlying planning system continues operating with outdated assumptions.

AI can monitor purchase-order history, demand variability, supplier reliability and inventory performance to identify where parameters no longer reflect operating reality. The proposed approach is incremental: start with bounded parameters, use existing ERP data and recommend changes before shortages, excess inventory or repeated expediting occur.

Operational perspective: parameter optimization should not become an uncontrolled process that writes directly into the ERP. Each parameter needs a business owner, an approved calculation method, minimum data-quality standards and a review cadence. Recommended changes should be tested against service, inventory, cost and capacity consequences before implementation.

Organizations should retain the current value, recommended value, supporting evidence, approver and effective date. High-impact changes—such as supplier lead times, safety-stock policies or minimum order quantities—also need manual overrides, exception thresholds and rollback procedures. A lightweight analytical layer can generate recommendations, while the ERP or APS remains the controlled system of record.