What the ERP already had, finally fed

The problem

The model is already built. What is missing is anybody feeding it.

Traceability modules built but empty, with the risk of mixing different product flows in the same plant.

• Almost the entire data model for end-to-end traceability already exists in your ERP: origin, animal, carcass, batch, order, logistics unit.
• What is missing is not software, it is the feeding, and every year a plant spends deciding which new system to buy is a year the model it owns stays empty.
• On top of that sits the specific risk of running several commercial flows through one plant: own production and toll orders that must never mix, on the same species, on the same day.
• A traceability incident in this business is not a data problem. It is product on the market you cannot bound, with a single impact that runs above €100,000 once you count the recall, the customer and the lost status.

How it fits the IRIS system

Five coordinated modules — origin, quality, changeover, condemnation, label.

Origin (certificate + ear tag) → Quality (real control) → Critical changeover (Edge + JIDOKA AI) → Condemnation (automatic dossier) → Shipping (GS1-128 label, no typing). SMED AI speeds up species/batch changeovers.

Origin and ear tag open the chain, quality control fills what the plant really checks, the changeover validation protects the boundary between species and clients, the condemnation dossier accounts for what leaves the chain, and the GS1-128 label closes it at the dock. SMED AI works on the changeovers themselves, so protecting the boundary does not cost throughput.

See the full IRIS architecture →

Before and after

The five modules, and what each one closes

Module	Which cases it draws on	What it closes
1 · Origin	Origin certificate and ear tag (7)	Which animal, from which holding, entered
2 · Quality	Quality module fed (6) + grading log (1)	What was checked, and on which carcass
3 · Changeover	Species changeover validated (10)	The boundary between species and between clients
4 · Condemnation	Condemnation dossier (11)	What left the chain, and where it went
5 · Shipping	GS1-128 label (8) + grading support (9)	Which box carries which batch, to whom
Cross-check	All five, against each other	The gap no single module can see

underused modules → real end-to-end traceability with the same ERP.

Impact estimate

Impact estimate — to be validated with your numbers.

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 6-12 months, counting only the modules already paid for that start being used.
• Protection against a traceability incident whose single impact runs above €100,000, before counting what it does to your standing with the customer.
• From underused modules to end-to-end traceability on the same ERP, with no new platform to buy and maintain.
• And the scope of any withdrawal moves from a whole week of production to the exact batch, which is the difference between an incident and a crisis.

Protection against a traceability incident (single impact > €100,000) and monetization of the investment already made in the ERP. Estimated payback of 6-12 months. *Estimate to be validated*.

And the fair question from the production manager

“Do we have to deploy all twelve cases before any of this works?” — no, and framing it that way is the surest way never to start. The modules are built in layers and each case pays on its own from the first month. With the grading log and the origin capture you already hold module 1 and half of module 2, which is where most of the missing traceability actually sits. Each layer rests on the same verification gate, an anchored extraction where the best models drop below 1.5% error [1] with a person confirming before anything posts.

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