Payoyo cheese — "which goat is this cheese from?" should not be a question without a concrete answer.

The problem

Five links, five systems, and a cheese that does not know which goat it came from.

The value of Payoyo cheese lies in the goat and in the pasture, not only in the aging. But between the animal and the cheese there are five links where information breaks:

1. The farm — REGA number, breed register, head count. In the farmer's notebook.
2. The milking — how much, what quality, which day. Sometimes on paper, sometimes in a notebook in the milking parlor.
3. The tanker collection — which farms the driver visited today, and in what order. A paper delivery note signed at every stop.
4. The dairy tank — which milk goes into which tank, blended with yesterday's. In the ERP, if there is one, or in a notebook.
5. The vat — which vat is made from which tank, at what time, in what quantity. On the master cheesemaker's sheet.

When a gourmet buyer, a journalist or a visitor asks "which goat is this cheese from?", the honest answer is "from the mountains". And the mountains are not a goat. Today, breed traceability per cheese is weeks of work after the fact, and almost nobody does it outside an audit. The cheese loses premium value because it cannot tell its origin with a scan.

How it fits the IRIS system

iLEAN Tracer joins the five links — without asking anyone to change what they already do.

The problem with breed traceability is not missing data. The data is there — in five different places. It is information that lives in islands and that, by the time it comes together, can no longer be tied to a specific cheese. iLEAN acts as the putty that fills the gaps between the farmer's notebook, the tanker, the ERP, the vat sheet and the cheese label — without forcing artisan farmers to digitalize overnight.

Tracer is an agent that lives on top of Connect. It captures the data where it already is (voice, photo, app or ERP), joins it by batch and leaves every cheese with a verifiable origin all the way back to the goat. The person signs — the chain builds itself from what is already there.

The iLEAN piece applied to Payoyo and breed traceability:

• iLEAN Connect at the farm — captures the day's milking through whichever channel each farm uses: a photo of the notebook, voice dictation into the earpiece, a WhatsApp to the driver, a simple app. Graded farm by farm: direct integration for those who have a system, assisted capture from a phone for those who do not. Nothing has to change.
• iLEAN Connect on the tanker — the driver logs every collection stop (which farm, how much, what time) with a couple of taps on the phone or by voice. The paper delivery note is still signed if it has to be — Connect reads it and cross-references it.
• iLEAN Connect at the dairy — captures which tank each tanker load goes into, which vat is made from each tank and which pieces come out of each vat. A modern ERP, a MES, Excel, the master cheesemaker's sheet: it all goes in.
• iLEAN Tracer (agent) — lives on top of Connect and joins the five links by batch. Every labeled cheese carries an identifier that returns, with a single scan, the list of farms of origin, the goat/sheep breeds and the collection day. Breed traceability per cheese, not per plant.

See the full IRIS architecture →

Before and after

Fragmented documentary traceability vs. breed traceability per cheese

Aspect	Classic control	With iLEAN Tracer + Connect
Origin of the milking	Farmer's notebook, never cross-referenced	Batch per farm, captured on the day
Tanker traceability	A delivery note signed at every stop	Full route with quantities, at second zero
Blending in the tank	Implicit, with no breakdown by origin	Tank composition by farm and by breed
Cheese ↔ breed	"From the mountains"	"From these three farms, 80% payoya / 20% grazalemeña"
Time to reconstruct	Weeks, during an audit	Available as the cheese leaves the vat
Burden on the farmer	Extra paperwork, resistance	Capture through his own channel (voice/photo/WhatsApp)

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

• Artisan dairy collecting from 5-10 payoya goat farms and 2-3 grazalemeña merino farms, one tanker of its own or shared, one vat room.
• Connect + Tracer pilot: rollout of voice/photo/WhatsApp capture for the driver and the master cheesemaker, light integration with the ERP or the master's sheet. First value expected within a few weeks: breed traceability per cheese available as the pilot batches leave the vat.
• Expected margin improvement on the gourmet/export channel of ≥ 30%, conservatively (the additional premium for cheese with a certifiable breed). Indicative payback between 4 and 9 months depending on volume, channel and average price per kg.
• The hard lever is the cheese sold with its origin told on the label: a premium price in gourmet and export, and the Regulatory Council's bonus for a creamery with accredited breed traceability.

And the master cheesemaker's reasonable doubt

"What if the AI mixes up a batch and puts down the wrong breed?" — 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 (the notebook entry, the tanker delivery note, the vat sheet), the best models brought error below 1.5% ^[1]. And even so, what is critical is never decided alone: Tracer proposes the traceability and the master cheesemaker signs the batch. The chain is always signed by a person.

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