Kanvaz AI Labs
All solutions

Decisions, not correlations

Causal AI

Understand what drives supply chain waste. Test interventions before changing your planning parameters.

  1. 01

    Trusted data

  2. 02

    Causal drivers

  3. 03

    Test interventions

  4. 04

    Planning feedback

How it works

01

Find the drivers

Separate consumption signals from distortions caused by batching, promotions and shortage ordering.

02

Compare interventions

Assess changes to planning policies and parameters against the causes of the problem.

03

Learn from outcomes

Measure the effect of approved actions and feed better parameters back into planning.

~/example

Stockouts and excess, together.

Find what drives inventory swings, compare interventions and validate their effect.

simulate: do(order batch = 4 weeks) · illustrative

upstream order swing

±32%

Test a parameter change on the causal model before you change it in planning.

Ask what happens if you change the policy.

Counterfactual · illustrative

  • Retail±8%
  • Distributor±12%
  • Manufacturer±18%
  • Supplier±26%

Order swing per tier under do(batch, promotion). Toy model for explanation, not client data.