Inventory Planning
Demand Is Not a Bell Curve
Your safety stock formula assumes demand is a neat bell curve. Real demand cannot go below zero. Where the formula breaks, and what to use instead.
Notes from the practice: graded forecasts, closed-network AI, and what industrial AI actually delivers, with every figure labeled by its source.
Inventory Planning
Your safety stock formula assumes demand is a neat bell curve. Real demand cannot go below zero. Where the formula breaks, and what to use instead.
Inventory Planning
The stockout cost most reports track is one lost sale. The real cost is spread across three other budgets that never get added to it. Where to look.
Demand Forecasting
A healthy accuracy score and the wrong items being wrong are not a contradiction. What forecast accuracy really measures, what it hides, and how to grade it.
Decision Analytics
A flat average can hide a business changing underneath it. Where averages mislead, when to use the median instead, and one check to run on your own numbers.
Inventory Planning
Most ERPs set a reorder point once, at setup, and never touch it again. Demand and lead time keep moving. Here is why that number quietly goes stale.
Business AI Transformation
AI implementation usually fails between the demo and the routine. What the data shows about that gap, and how staged structure with fixed KPIs closes it.
Secure & Sovereign AI
Microsoft runs a disconnected GPT-4 for US intelligence; France runs self-hosted army AI; the US Army serves ~75,000 users. Sovereignty is a requirement.
Demand Forecasting
More Retail took forecast accuracy from 24% to 76%; Walgreens cut over-forecasting from 15% to 1%. The common thread is not the algorithm: someone graded it.
Manufacturing AI
Verified, source-labeled results from Audi, BMW, a peer-reviewed study and Israeli manufacturer Kanfit: what industrial AI really delivers in production.