Manufacturing

Manufacturers sit on rich process and supply data, but conventional analytics cannot untangle which factors genuinely drive cost, quality and risk. Causal AI can, and it changes what planners and engineers are able to act on.

Causal MRP: eliminating $55M of annual premium freight

Global agricultural equipment manufacturer

$55M annual reduction in premium freight costs, delivered alongside existing SAP MRP with no platform replacement

Challenge

Premium freight costs were running at $165M per year, driven by reactive ordering triggered by downstream supply shortages. Existing SAP MRP was optimized for steady-state conditions and could not distinguish supply events that required urgent action from those that would resolve without intervention.

Approach

A Causal MRP solution was built alongside the existing SAP deployment, adding a causal reasoning layer that identified the true drivers of stock-out risk and premium ordering events. The model separated genuine causal risks from correlated noise, so supply planners could intervene earlier and with greater precision.

Outcome

Premium freight costs reduced by $55M per year. Purchasing interventions became better targeted and better timed, and planners could justify every corrective action with a clear causal rationale rather than threshold-based alerts from an opaque system.

Causal AIDeterministic CodeML/DL

Electrolysis optimization: improving anode quality in aluminum production

Global metal manufacturer

Challenge

Carbon anode quality directly affected energy consumption, production cost and emissions in aluminum electrolysis. Consistent anode resistivity was commercially critical, but the interactions between production parameters and quality outcomes were poorly understood.

Approach

A causal model of the anode production process was built with resistivity as the primary outcome, incorporating parameters including coke granulometry, pitch percentage and mixer speed. Historical production data quantified the causal relationship between each input and the quality of the finished anode.

Outcome

Average anode resistivity improved substantially with significantly fewer outliers. Lower resistivity directly reduced energy consumption per ton of aluminum produced, improving process economics and cutting emissions across the facility.

Causal AIKnowledge GraphDeterministic Code

Production quality intelligence: reducing defect rates in precision manufacturing

Global precision components manufacturer

Challenge

Defect rates could not be attributed to any single cause through conventional analysis. Process parameters, material inputs, equipment state and environmental conditions all contributed to quality variation, but the interactions between them were too complex to untangle manually.

Approach

A causal model of the production process connected equipment telemetry, material input quality, environmental conditions and output quality data. It identified the specific causal combinations driving defect rates, enabling targeted process corrections rather than broad parameter tightening that reduced throughput.

Outcome

Defect rates reduced materially once the causal process drivers were identified and corrected. The manufacturer gained a repeatable diagnostic framework for quality variation and an evidence-based rationale for every process change investment.

Causal AIML/DLDeterministic Code

The right problem. The right AI. Working in weeks.

If you recognize your challenge here, or want to discuss one that is not covered, we would like to talk.

Book a discovery session See a demo