RIBER traceability for Iberian ham — tying the dehesa to the seal, without the batch close costing a week of spreadsheets.

The problem

RIBER exists — keeping it clean without extra manual work does not.

Iberian pig traceability crosses at least six links and five different systems, almost never connected to each other:

1. Dehesa and livestock — ear tag, batch, entry weight into montanera. SIGGAN, the farm logbook, RIBER.
2. Montanera and finishing farm — feeding, ITSA. A PDF from the certification body, paperwork from the finishing farm.
3. Slaughter — carcass number, classification. The slaughterhouse MES (sometimes in house, sometimes a third party's).
4. Cutting — the cut extracted, the barcode. MES or a thermal printer label.
5. Drying house — entry into curing, chamber control, exit. ERP, chamber sheets, the master ham curer's judgment.
6. Sealing and sales — color, seal number, shipping. A scanner gun and a sticker.

RIBER is the official source of truth. Whatever does not appear in it fills parallel spreadsheets that two people maintain every month to close the accounting batch. When the certification body or the Ministry of Agriculture asks for piece-by-piece traceability, that week of spreadsheets turns into two. And the day the cutting room lead retires, part of the chain leaves with them.

How it fits the IRIS system

iLEAN is not a seventh system — it is the putty that stitches together the RIBER, the MES and the ERP you already have.

The problem with Iberian traceability is not a lack of data: it is data living on islands (RIBER, ITSA, MES, ERP, the logbook, the vet's judgment) that only gets cross-referenced by hand at closing. iLEAN Tracer acts as the putty that fills those gaps, without asking you to change your RIBER, your MES or your ERP.

Connect captures every link where it happens. Edge watches the critical points (yard, cutting, seal). The agent stitches the chain and blocks the close if something does not add up. The person signs — never the other way round.

The iLEAN pieces applied to RIBER traceability from the dehesa to the seal:

• Connect — captures RIBER, SIGGAN, the certification body's ITSA, the slaughterhouse/cutting MES and the drying house ERP. It also captures the voice note dictated by the cutting room lead, the photo of the ear tag in the yard, the email from the certifier with a change of criteria. Bidirectional: it also drafts the replies to the certification body, which a person signs.
• Edge — AI vision terminals at the three critical points: the handling yard (reads the ear tag and the traits), the cutting room (reads the carcass mark and the cut extracted), the sealing line (reads the seal color and the piece). Each terminal runs locally: it does not depend on the network to record.
• Agent — the brain that stitches. It lives in Central, cross-references links as they happen, builds the batch file live (not at closing) and blocks progress if the chain breaks. If the finishing farm's ITSA is missing, it says so on the day it happens, not three months later.

See the full IRIS architecture →

Before and after

Closing the batch by hand vs. traceability stitched with iLEAN Tracer

Aspect	Traditional traceability	With iLEAN Tracer (Connect + Edge + Agent)
Capture per link	Each one in its own system or on paper	Unified at second zero, anchored to ear tag/batch/piece
Batch close	A week of spreadsheets every month	The file is ready the day the batch closes
Gap detection	External audit or a complaint	The day the data is missing — with room to fix it
Vet's / master ham curer's knowledge	In their head, gone at retirement	Captured as model criteria (with their signature)
Operation with no network on the farm	n/a	Edge keeps recording locally and syncs when it is back
Certification body (ENAC) and Ministry audits	Manual rebuild, weeks of work	Piece-by-piece dossier in one click

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 chain. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.

• Iberian group with its own livestock, slaughterhouse/cutting and drying house, several thousand pieces a year, acorn-fed and field-fed certifications managed with a logbook + ERP + ITSA reports as PDFs.
• Tracer pilot on two critical links (handling yard ↔ slaughterhouse and cutting ↔ sealing). First value expected within a few weeks: the batch file goes from being closed a week later to being ready the day it closes.
• Indicative payback between 4 and 9 months, depending on your current batch-closing hours, how often the certification body raises nonconformities, and the cost of the recall risk from incomplete traceability.
• Expected reduction in manual closing work ≥ 30%; the rest translates into fewer arguments with the certifier and a better position in a Ministry inspection.

And the quality manager's reasonable doubt

“What if the AI gets the stitching wrong and lets a weak link through?” — hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI merely recontextualizes a piece of data from one system into another (reading the ITSA, tying it to the ear tag, checking it against RIBER), the best models brought error below 1.5% ^[1]. And even so, what is critical is never decided alone: the agent blocks the close 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.