Casein mark traceability per wheel in PDO Parmigiano Reggiano — a dossier per wheel, not an Excel per shift.

The problem

A wheel with two years in the cave behind it cannot be improvised.

PDO Parmigiano Reggiano is worth what it is worth precisely because its traceability is total: every wheel, identified by its casein mark, has to be linkable to the morning and evening milk, the vat pH, the rennet, the thermocouples, the ripening room and the Consorzio's final inspection. Two years later.

In most caseifici that lives in three separate places:

1. The vat paperwork — the casaro's notebook or Excel file, with the day's parameters and the handwritten codes.
2. The ERP — milk intakes, supplier batches, warehouse issues.
3. The Consorzio portal — the official record that decides whether the wheel carries the PDO seal once it passes inspection.

And the casaro, who carries the map in his head, cannot be in all three places at once. When a wheel travels from the vat to the brining room, from there to the cave, and two years later into the Consorzio file, one badly copied notebook entry is enough to break the link — and that is half a wheel gone from the year's balance sheet.

How it fits the IRIS system

iLEAN Tracer does not add a fourth system — it seals the cracks between the three you already have.

The problem with per-wheel traceability is not a lack of information: it is information that lives in islands and that, at the critical moment (the Consorzio inspection, an importer's query, a complaint), does not reach the person who decides in time. iLEAN acts as the putty that fills those gaps, without asking you to change your ERP, your vat system or the Consorzio portal.

Edge reads the casein mark at the vat outfeed. Connect captures the casaro's paperwork and the Consorzio portal. The agent cross-references, holds the batch if something does not add up and prepares the per-wheel dossier. The person signs — never the other way round.

The three iLEAN pieces applied to the casein mark per wheel:

• Edge — a machine-vision camera (CNN) on the vat outfeed cart, at the entrance and exit of every cave and at the final branding. It reads the alphanumeric code on the casein plate in milliseconds. It works with no network: down in the cave, where Wi-Fi never reaches, Edge keeps reading and storing.
• Connect — captures the casaro's paperwork (a photo of the notebook or a reading from the vertical MES), the ERP data and the feed from the Consorzio portal. And it also captures what comes in from outside (an importer's email with a new Consorzio rule, a WhatsApp from the cave manager with an update) at second zero, without anyone having to forward anything.
• Agent — cross-references the mark, the milk batch, the vat parameters, the cave record and the Consorzio file. If a wheel does not match its file, it does not send an email at 10 p.m.: it flags it and alerts the casaro through whichever channel he uses. The person validates and signs; a wheel never moves out of the record on its own.

See the full IRIS architecture →

Before and after

Manual traceability vs. cross-referenced traceability with iLEAN Tracer

Aspect	Vat notebook + Excel	With iLEAN Tracer (Edge + Connect + Agent)
Reading the casein mark	The casaro writes it down by hand at the vat outfeed	An Edge camera reads it in milliseconds and links it
Link to the milk batch	Through the vat sheet — recovered by searching	Automatic, at second zero
Cave record	Physical notebook, risk of loss	A reading on entering and leaving every cave
Cross-check against the Consorzio file	Manual, whenever the inspector asks	Dossier per wheel, one click away
Operation with no network	n/a (paper)	Edge keeps running underground on the panel's own light
An importer's complaint	Rebuilt by hand, days	An answer with photo and dossier, in hours

Impact estimate

Impact estimate for your plant — to be validated with your numbers.

The block below is an estimate to be validated with the specific data of your caseificio. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.

• PDO Parmigiano Reggiano caseificio, output of ~80-150 wheels/day, 2 ripening caves, an active Consorzio file.
• Edge pilot on the vat outfeed cart and at the entrance of one cave, with integration of the casaro's paperwork and the Consorzio portal. First value expected within a few weeks.
• Indicative payback between 4 and 9 months, depending on the frequency of file incidents and the average value of a wheel recovered.
• Expected reduction of ≥ 30% in the hours spent rebuilding files by hand, with a per-wheel dossier ready for the Consorzio inspector or the importer.
• The hard lever is a single PDO wheel rescued from a traceability failure: given what a wheel matured for two years is worth, one a year pays for the pilot.

And the casaro's reasonable doubt

"What if the AI misreads the mark and links a wheel wrongly?" — hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI simply recontextualizes a piece of data from one system into another (reading the plate, tying it to the vat record), the best models brought error below 1.5% ^[1]. And even so, what is critical is never decided alone: iLEAN holds the doubtful wheel and the person signs. The three safety rings exist precisely for this.

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