The raw-material delivery note, straight to the ERP by photo

The problem

The truck waits while the delivery note is typed line by line.

The carrier waits racking up demurrage while the receiving lead types. If they rush, batches get mistraced and surface weeks later when a line produces out of spec.

• Clay arrives in bulk, feldspar and sand by truck, frits and pigments in big bags or sacks. Each delivery has a note with several batches and quantities. Each supplier prints it differently. Some send the batch only on the bag label, others only on the note, and the receiving lead has to reconcile both by hand.
• The receiving lead types it line by line into the ERP while the carrier waits and demurrage adds up. On a busy morning, two or three trucks queue at the gate. Every minute is charged by the carrier and none of it adds value.
• When they rush, a batch is assigned to the wrong silo or the wrong material, and nobody notices at the time. The receipt looks correct in the ERP; only the physical batch is somewhere else.
• The error surfaces weeks later, when a line produces out of spec and the traceability points to a batch that was never in that silo. By then the batch is in the mill, the press and the kiln. Finding which tiles carry it takes days.

How it fits the IRIS system

Connect in photo mode — the delivery note and the labels, read in seconds.

Connect photo mode. AI vision recognizes the delivery note's layout and the labels (GS1-128, QR, barcode, plain OCR). One-tap validation on an industrial tablet.

The receiving lead photographs the delivery note and the batch labels, checks a summary on the tablet and taps once. The carrier leaves in minutes and every batch is in the ERP with its silo or location from the start. Traceability from the finished tile back to the supplier stops having holes. And the receiving lead spends the morning checking material, not typing.

See the full IRIS architecture →

Before and after

Receiving clay and frits today versus receiving by photo

Aspect	Today	With iLEAN Connect
Truck time at the raw-material gate	Long, with demurrage	Minutes
Typing the delivery note	Line by line	One photo and one tap
Batch labels on big bags	Copied by hand	Read via GS1-128, QR, barcode or OCR
Batch assigned to the wrong silo	Found weeks later	Caught at the validation tap
Traceability from line back to supplier	Broken in places	Complete
Warehouse users in the ERP	One per person typing	Fewer named users

Long carrier delay → minutes. Hidden batch errors → zero. Broken traceability → complete via photo and GS1.

Impact estimate

Impact estimate — to be validated with your numbers.

The block below is an estimate to be validated against your plant's actual data. We put it forward so the committee has an order of magnitude; we refine it during the assessment.

• Estimated payback 3-8 months.
• CFO angle: a typical 50-70% reduction in receiving and warehouse-module ERP licenses.
• Carrier delay goes from long waits to minutes, and demurrage stops adding up at the gate. Carriers notice, and the plant stops being the slow stop on their route.
• Hidden batch errors drop to zero, and traceability from the line back to the supplier is complete via photo and GS1. When a line goes out of spec, the batch it was fed from is known in one query.

CFO angle — typical 50-70% reduction in receiving/warehouse-module ERP licenses. Estimated payback 3-8 months. *Estimate to validate*.

And the fair question from the production manager

“What if it reads a batch number wrong?” — reading a delivery note with a known layout and labels with a barcode or QR is an anchored task, where the best models drop below 1.5% error [1]. And the receiving lead sees the summary on the tablet before the tap: a batch that does not match the purchase order is flagged, not pushed. The same happens with a silo that does not accept that material.

^[1] OpenAI paper "Why Language Models Hallucinate", 2025 — on the reliability of AI in anchored tasks.