Origin docs and ear tag enter on their own

The problem

The first record of the animal is the slowest one to create.

Typing the certificate by hand is slow and error-prone; the ear-tag record simply isn't filled in today.

• The origin certificate carries fifteen or more fields per consignment: holding, species, count, transport, dates, health statements, signatures.
• It is typed by hand while the truck waits, which is the worst possible combination of pressure and detail.
• The ear-tag field exists in the system and is simply not filled, because reading tags one by one and typing them is not something a reception shift can absorb.
• And an empty ear-tag field is not a cosmetic gap: it is the difference between traceability down to the animal and traceability down to the truck. Every claim made further downstream about where a piece of meat came from rests on that field, and today it rests on nothing.

How it fits the IRIS system

Connect in photo mode — one photo of the document, one of the tag.

Connect photo mode. Recognizes the certificate and the tag, extracts the fields and, after validation, inserts them into the ERP.

The certificate and the tag are recognized as what they are, the fields are extracted and, once validated on screen, inserted into the records that already exist in your system. Nothing new is built and no parallel register appears. Reception stops being the bottleneck of the traceability chain and becomes its first reliable link, which is where every later question ends up pointing.

See the full IRIS architecture →

Before and after

Reception today versus reception captured

Aspect	Today	With iLEAN Connect
Certificate fields	Typed one by one	Extracted and validated
Ear-tag record	Empty	Filled on every reception
Who does the entry	Two or three people with licenses	The lead, validating
The truck at the dock	Waits for the typing	Waits for the photo
Traceability reaches	The consignment	The animal
A doubtful field	Typed as best guess	Flagged before it is stored

empty ear-tag table → populated on every reception.

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 3-8 months, depending on receptions per week and species mix.
• The receiving and traceability module typically needs two or three named licenses; with capture only the lead validates, a typical 50-70% reduction.
• From an ear-tag table that has never been populated to one filled on every reception, which is what makes animal-level traceability real rather than declared in a procedure nobody can evidence.
• And the certificate stops being retyped: the original image stays attached to the record, so whoever checks it later reads the source instead of somebody's transcription of it.

The receiving/traceability module typically requires 2-3 licenses; with Connect only the lead validates — typical 50-70% reduction. Estimated payback of 3-8 months. *Estimate to be validated*.

And the fair question from the production manager

“An ear tag is small, dirty and moving — can it really be read?” — reading a tag is an anchored task with a known format and a checkable structure, where the best models drop below 1.5% error [1]. There is a second net: a tag whose number does not fit the expected structure, or whose reading is uncertain, is not stored as a fact. It comes back for the operator to confirm, and the photo stays attached either way.

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