Back to Portfolio
AI/Automation 2024 8 months

Inventory & Supply Chain AI Optimization

Manufacturing Conglomerate

A manufacturer was losing $3M a year to stockouts and overstock because planning ran on spreadsheets and gut feel. Bricksense built an AI forecasting and replenishment layer on top of its existing ERP.

Demand ForecastingML EngineeringInventory OptimizationSystems IntegrationAnalytics
↓72%
Inventory Waste
94%
Forecast Accuracy
$2.8M
Annual Savings
+55%
Order Fulfillment

The Challenge

Demand planning was manual, monthly, and inaccurate. Fast-moving lines stocked out while slow ones tied up cash and warehouse space. Nobody trusted the numbers enough to automate against them.

The solution had to work with the incumbent ERP, handle thousands of SKUs across multiple plants, and be explainable enough for planners to act on.

Our Approach

1

Forecast models

Per-SKU demand models blending seasonality, trend, promotions, and external signals, backtested against history.

2

Inventory policy

Safety-stock and reorder points optimised per SKU and location against service-level targets and holding cost.

3

ERP integration

Forecasts and recommended orders write back into the ERP planning screens planners already use.

4

Explainability

Every recommendation shows its drivers and confidence, so planners can override with reason codes that feed back into the model.

The Solution

Planners now start from a per-SKU forecast and a recommended order they can accept or adjust, rather than a blank spreadsheet. The model learns from their overrides.

Inventory is rebalanced toward service levels the business actually set, freeing cash from dead stock while cutting stockouts on the lines that matter.

What We Delivered

  • Per-SKU demand forecasting models with backtesting
  • Inventory optimisation engine (safety stock, reorder points)
  • Write-back integration with the existing ERP
  • Explainable recommendations with override capture
  • Planning analytics and accuracy dashboards

The Results

Forecast accuracy reached 94%, cutting inventory waste by 72% and improving on-time order fulfilment by 55%.

The combined effect of less dead stock and fewer stockouts returned $2.8M a year, close to eliminating the original loss.

↓72%
Inventory Waste
94%
Forecast Accuracy
$2.8M
Annual Savings
+55%
Order Fulfillment
Our planners trust the forecast now because they can see why it says what it says. We freed up cash and stopped stocking out on our best sellers.
VP of Supply ChainManufacturing Conglomerate