The delivery note that no longer blocks the truck

The problem

Typing the delivery note line by line keeps the truck waiting at the dock.

If the receiving lead rushes to free up the truck, the lot ends up poorly traced — showing up weeks later as an origin non- conformance nobody can track back to the exact casting lot.

• Critical raw material arrives every day: recycled lead ingots, sulfuric acid, PE separators, containers and lids, each with its own delivery note format. Formats change by supplier and sometimes by shipment, so no template survives for long.
• The receiving lead has to type it line by line into the ERP while the truck sits at the dock for around 45 minutes, racking up demurrage. Meanwhile the next truck waits in line, and the dock becomes a bottleneck by mid-morning.
• If he rushes to free the truck, the lot gets recorded loosely: a supplier lot number skipped, a quantity rounded. Nobody does it on purpose; it is the predictable result of typing under time pressure.
• Weeks later that shows up as an origin non-conformance nobody can trace back to the exact casting lot that used the ingot. The investigation then stops at the dock, because the record that should link them was never complete.

How it fits the IRIS system

Connect in photo mode — the supplier's delivery note, read as it comes.

Connect photo mode. Vision AI recognizes the supplier's own delivery-note format as-is and extracts the tabular structure; for pallet labels it reads GS1-128, QR, barcode or plain OCR. Validation with a single tablet tap.

Every lead and acid supplier keeps its own delivery note. iLEAN does not ask them to change anything: it reads each format as it is, adds the GS1-128 or QR on the pallet label, and the receiving lead only confirms. The truck leaves sooner and the lot enters the ERP complete, which used to be a trade-off and no longer is.

See the full IRIS architecture →

Before and after

Battery raw material receiving today versus by photo

Aspect	Today	With iLEAN Connect
Truck time at the receiving dock	Around 45 minutes	A few minutes
Recycled lead lot number	Sometimes skipped in the rush	Read from the note and the pallet label
Pallet label formats	Typed by eye	GS1-128, QR, barcode or plain OCR
Link from ingot lot to casting lot	Broken or incomplete	Complete, by photo and GS1
Hidden lot errors in the ERP	Found weeks later	Zero
Receiving and warehouse module seats	One per person who types	Only the validator

Truck blocked ~45 min → minutes. Hidden lot errors → zero. Broken origin traceability → complete, by photo and GS1.

Impact estimate

Impact estimate — to be validated against your ERP invoice.

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.

• 💰 CFO angle: typically a 50-70% reduction in receiving and warehouse module licenses, recurring every year.
• Estimated payback 3-8 months. The saving is measured on your own dock times before anything is committed.
• Truck time at the dock from around 45 minutes to a few minutes, with less demurrage. Fewer demurrage charges and a dock that no longer backs up by mid-morning.
• And origin traceability from recycled lead ingot to grid casting lot becomes complete, instead of broken in the rush. That is what lets an investigation follow a defect back to the ingot.

CFO angle — typical 50-70% reduction in receiving/warehouse- module licenses, estimated 3-8 month payback. *Figure to validate*.

And the fair question from the production manager

“Every lead supplier sends a different delivery note — will it read them all?” — recognizing the table structure of a known document and the codes on a pallet label is an anchored task, where the best models drop below 1.5% error [1]. And the receiving lead confirms the summary with one tap before the lot enters the ERP, so a doubtful field never goes in unchecked. The original photo is stored with the lot as evidence.

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