Details

Back to Adopt AI

Part of

Adopt AI

Use case

Inventory Planning

Forecasting and AI recommendations support inventory orders and transfers with human approval.

  • Problem

    The customer needed to decide what to reorder or transfer for the next month based on current stock, historical sales and manufacturing/warehouse constraints.

  • My role

    I worked on an anonymized forecasting and approval workflow that connected inventory systems, local forecasting jobs, LLM recommendations, human approval and business writebacks.

  • Implementation

    Integrate inventory, stock, warehouse and manufacturing-order systems; run forecasting models inside RCE sandboxes; let LLM agents interpret history and forecasts; require human approval; then execute approved orders or warehouse transfers.

  • Engineering choices

    Keep forecasting and recommendations advisory until a human approves; only approved orders or transfers become effectful writebacks.

Stack and domains

  • Python
  • ML forecasting
  • LLM agents
  • RCE
  • enterprise integrations