Photo to the delivery note and the resin octabin label

The problem

Every octabin batch is typed by hand while the truck waits.

A truck arrives with resin octabins. The receiving supervisor has to type batch after batch into the ERP while the truck racks up billable demurrage. If he rushes, batches get entered wrong and show up weeks later when a customer complains and nobody can trace which octabin produced which closure.

• A truck arrives with resin octabins. The receiving supervisor has to type batch after batch into the ERP: supplier, grade, resin batch, weight per octabin. With twenty or more octabins per truck, the typing alone takes most of the unload.
• Meanwhile the truck sits at the dock and demurrage starts to bill. At around 45 minutes per unload, receiving becomes a bottleneck on busy days.
• If he rushes, a batch number gets entered wrong. Nobody notices until weeks later, when a bottler complains and no one can trace which octabin produced which closures.
• The receiving module needs 2-3 active ERP licenses just for this data entry, paid every month whether trucks come or not. Those licenses exist for typing, not for deciding.

How it fits the IRIS system

Connect in photo-to-analog mode — delivery note and octabin label, one tap.

With Connect's photo-to-analog mode, he photographs the delivery note and the octabin label. Vision AI extracts batch, quantity, supplier and resin reference. The supervisor validates with one tap on the tablet and Connect pushes everything to the ERP via API — truck unloading no longer blocks him.

The supervisor stops being a typist. He photographs the delivery note and the octabin label, checks a summary on the tablet and confirms: resin batch, quantity, supplier and reference enter the ERP through the API. The truck leaves, and the batch is traceable down to the closure. Receiving stops being the weakest link of the closure plant's traceability. And the supervisor goes back to checking material, not typing it. Checking the octabins for damage or moisture is the part of receiving that actually needs his eyes.

See the full IRIS architecture →

Before and after

Resin receiving today versus receiving by photo

Aspect	Today	With iLEAN Connect
Resin truck at the dock	About 45 min	About 5 min
Resin batch entry	Typed octabin by octabin	Extracted from the label
Batch typo at receiving	Found when a bottler complains	Caught at the one-tap check
Octabin-to-closure traceability	Broken by entry errors	Batch linked from receiving
ERP receiving licenses for octabin entry	2-3 active users	−50 to −70% typical
Resin supplier documents	—	Unchanged, any format

Truck idle 45 min → 5 min; -50/-70% receiving-module licenses.

Impact estimate

Impact estimate — to be validated with your resin deliveries per week.

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, combining licenses and dock time. Both savings start with the first truck.
• A typical 50-70% reduction in receiving-module licenses: today 2-3 licenses are active for entry; with Connect only the supervisor validates. The rest of the team keeps read access where it needs it.
• Truck time at the dock drops from 45 minutes to about 5, which cuts billable demurrage and frees the dock for the next delivery. On days with several trucks, that is the difference between a queue and a flow.
• Correct resin batches from the first minute protect the trace from octabin to closure when a claim arrives, which is the question that matters most to a bottler. A complaint can be traced back to the octabin in minutes instead of days.

Payback 3-8 months · −50 to −70% receiving-module licenses. The receiving module typically requires 2-3 active licenses; with Connect, only the supervisor validates. Typical 50-70% reduction in receiving-module licenses. Estimated payback 3-8 months. Estimate to validate.

And the fair question from the production manager

“Every resin supplier prints a different delivery note and label — will it cope?” — yes, without asking any supplier to change anything. Reading the table of a delivery note and the batch on an octabin label is an anchored task, where the best models drop below 1.5% error [1]; barcodes are read where they exist and printed text where they do not. And the supervisor validates with one tap before anything enters the ERP. A delivery with a doubtful field never enters silently.

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