Forecast Demand Amplification Index (FDAI)

GMI-MIN-029MiningPilot-Ready

Definition

Ratio of order variance to demand variance for AI-forecasted parts categories, measuring whether the AI ordering pattern amplifies or dampens supply chain signals.

The Measurement Gap

The Measurement Gap

When AI takes over parts forecasting, it processes predictive maintenance alerts, consumption history, and lead time data to generate purchase orders. In theory, this should produce smoother ordering patterns because the AI can anticipate demand rather than reacting to stockouts. In practice, AI forecasting systems often amplify demand signals rather than dampen them.

The amplification happens because predictive models react to the same signals simultaneously. When a fleet-wide condition triggers multiple predictive alerts, the AI may spike orders for a parts category across all affected assets at once. The AI is not wrong about the need, but the ordering pattern concentrates demand into a narrow window rather than distributing it over time. Suppliers see a surge, followed by a lull, followed by another surge. This volatility degrades supplier reliability, extends lead times, and can trigger exactly the stockouts the AI was trying to prevent.

Traditional procurement metrics track fill rate and stockout frequency but do not measure ordering pattern volatility. A procurement team could maintain 95% fill rate while generating wildly volatile order patterns that systematically degrade supplier relationships. The supplier's lead time increases are then attributed to "supplier performance" rather than to the ordering pattern that caused them. Without FDAI, organizations cannot detect whether AI-driven procurement is creating the supply chain instability it was designed to eliminate.

Decision Supported:

Is the AI parts forecasting system creating a more stable or more volatile ordering pattern than the demand it is predicting, and is that volatility degrading supplier reliability?

Formula

Plain Language: For each AI-forecasted parts category, calculate the variance of order quantities placed per period over the measurement window. Then calculate the variance of actual consumption quantities per period over the same window. Divide order variance by consumption variance. A result above 1.0 means the AI's ordering is more volatile than actual demand. A result below 1.0 means the AI is smoothing demand signals.

Notation:

FDAI = Var(Orders_AI) / Var(Consumption_actual)

Variable Description Source
Var(Orders_AI) Variance of AI-generated order quantities per period (typically monthly) for a specific parts category. Use population or sample variance consistently across all categories. ERP purchase order records filtered to AI-generated orders
Var(Consumption_actual) Variance of actual parts consumption per period for the same parts category. Consumption is measured as parts issued from warehouse to work orders, not parts ordered. CMMS/EAM warehouse issue records or inventory management system

Proposed formula, pending validation pilots.

At a Glance

Unit
Ratio
Trend Direction
Toward 1.0; Approaching 1.0 from above
Reporting Frequency
Monthly
Category
Supply Chain Intelligence Effectiveness
← Consequence-Weighted Stockout Exposure (CWSE)Emergency Ventilation Response Latency (EVRL) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment