PDO Dehesa de Extremadura — bellota-cebo traceability per piece, not per aggregated batch.

The problem

A piece is only worth what its whole chain proves — and the chain lives in pieces across four systems.

PDO Dehesa de Extremadura does not certify a batch: it certifies one specific piece. And to certify it, it demands to know, without ambiguity:

1. Which dehesa the pig came from — the montanera census, inside the protected area.
2. What it ate — bellota, cebo de campo or cebo, with weigh-in and weigh-out for the montanera.
3. What breed it was — a 100% Iberian or 75/50% breed certificate, with the chip linked to the animal.
4. How it was cured — drying room and cellar inside the PDO area, with a controlled curve and the Council's seal.

Every stage is recorded somewhere different: the census in the Regulatory Council's database, the scale on the farm, the breed certificate on a sheet from the vet, the drying room in SCADA, the seal in the ERP. When the auditor turns up and asks about piece X, somebody has to rebuild the entire chain by hand — and sometimes you find out a link is missing when it can no longer be fixed. The piece falls out of PDO or, worse, stays in with incomplete evidence and the Council finds it on the next visit.

How it fits the IRIS system

iLEAN Tracer captures every stage at source — the chain is not rebuilt, it is built live.

The PDO traceability problem is not missing data: there is data at every stage. The problem is that the data lives in islands and only gets assembled by hand, too late. iLEAN Tracer acts as the putty that fills the gaps, without asking you to change your ERP, your SCADA or the Council's database.

Connect captures everything from the dehesa to the seal. Edge reads the chip and the stamp in the plant. The agents cross-reference the PDO specification and build the per-piece dossier continuously. The person signs — never the other way round.

The iLEAN pieces applied to PDO Dehesa de Extremadura:

• Connect — captures the dehesa census during the montanera, the weigh-in/weigh-out from the stockman's scale, the vet's breed certificate (photo or app), the curve from the drying room SCADA, the cellar readings, the seal from the ERP. And it captures what comes from outside at second zero: an email from the Regulatory Council with a change to the specification, a message from the US importer with a new FDA requirement, a report from the stockman about an incident during the montanera.
• Edge — a terminal with vision and RFID reading on the farm scale and at the cellar sealing station. It reads the chip on the animal and on the piece, and associates photo, weight and PDO reference at second zero. With no network, it keeps reading.
• Agents — cross-reference everything with the PDO Dehesa de Extremadura specification and, in parallel, with the importer's requirements (US FDA, China GACC, premium EU retailer). If a piece has a broken chain, they raise the flag before sealing. If a piece already certified as PDO needs to ship to the USA, the dossier is recontextualized into the FDA format without anyone copying a field.

See the full IRIS architecture →

Before and after

Traceability rebuilt by hand vs. continuous traceability with iLEAN Tracer

Aspect	Classic PDO compliance	With iLEAN Tracer
Traceable unit	Aggregated batch, with interpolation	Individual piece, by chip and seal
Montanera census	The stockman's paper sheet	Connect captures it at second zero
Montanera weigh-in/weigh-out	The scale logbook	Edge reads chip + scale and links them
Broken chain before sealing	Found during the audit	The agent raises the flag
Dossier for FDA / GACC / US retailer	Manual reformatting, weeks	Recontextualized automatically
Quality team hours per audit	Weeks	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 plant. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.

• A PDO Dehesa de Extremadura plant with its own or contracted dehesas, a natural drying room and cellar, x pieces/year under PDO, part of the production exported to the USA or China.
• Tracer pilot with Connect out on the dehesa (scale + census + breed certificate), Edge at the sealing station and agents that build the per-piece dossier with PDO + FDA + GACC outputs. First value within a few weeks.
• Expected reduction in quality-team hours per audit above 30% against the baseline. Expected reduction in pieces that fall out of PDO because of incomplete documentation, above 30%.
• Indicative payback between 4 and 9 months, depending on PDO volume per year and the share of production exported with a per-piece dossier.
• Hard levers: one piece recovered for PDO that would previously have dropped out (bellota PDO price >> cebo price) — that pays for weeks of pilot. A single export pallet return avoided — that pays for the entire pilot.

And the CAIO's reasonable doubt

“What if the AI invents evidence that is not there?” — hallucination is a problem of free generation, not of anchored tasks. Here the AI only recontextualizes real data (the chip reading, the SCADA curve, the photo of the seal) into the PDO specification format or the FDA format. In tasks anchored to the source, the best models brought error below 1.5% ^[1]. And even so, what is critical is never decided alone: the agent prepares, the quality team 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.