THE CHALLENGE
-
Manual counting across 10+ product categories
Staff visually identified and tallied crates across 10+ categories on a moving conveyor - accuracy dropped significantly during high-volume cycles.
-
No real-time SAP update
Manual entry after cycle close created a lag between physical dispatch and system record.
-
No per-cycle audit trail
No reliable record of category-wise crate count per outbound cycle - making dispute resolution and reconciliation difficult.
-
Human error in category differentiation
Similar-looking crates across categories led to consistent miscategorisation under time pressure and high volumes.
THE SOLUTION
We deployed an AI-powered conveyor belt counting system that automatically detects, classifies, and counts outbound crates by product category in real time as they move through the dispatch line. Counts are pushed directly to SAP ERP at the close of every outbound cycle - eliminating manual entry entirely and keeping inventory records in sync with physical dispatch.
Flow: Crates move on conveyor -> AI detects and classifies each crate by product category -> Per-category count updated in real time -> Cycle closed -> Count pushed to SAP ERP -> Full cycle log saved to dashboard
-
Conveyor belt camera feed integrated with existing infrastructure to capture every crate passing through the outbound dispatch line.
-
AI model trained to classify crates across 10+ product categories - distinguishing between categories accurately even at conveyor speed and under varied lighting.
-
Per-category count maintained separately in real time throughout each outbound cycle, with running totals visible on the dashboard.
-
Automated SAP ERP integration pushes confirmed category-wise counts at cycle close - no manual data entry required.
-
Full per-cycle audit trail logged with timestamp, category-wise count breakdown, and images for every outbound cycle.





