Table olive traceability with AI — from the grove to the shelf, without rounding up spreadsheets the night before the audit.

The problem

The batch lives in three worlds that hardly ever talk to each other.

A table olive crosses three worlds before it reaches the shelf, and each one keeps its share of the truth on a separate island:

1. The grove and the dock — variety, parcel, harvest date, weight, on a delivery note the carrier hands over on paper. The dock operator writes it into a notebook or a spreadsheet.
2. Fermentation — tank, brine, date in, fermentation time, the NaOH batch used for the lye treatment in Spanish-style processing. In the cellar supervisor's notebook, or in a vertical MES nobody else in the plant ever opens.
3. Packing — size grade, format, packing date, heat treatment, printed code. In the ERP, in the line MES, or in whatever the inkjet prints at that moment.

As long as everything is fine, nobody asks. The day a retailer wants a dossier within 24h, or the night before an IFS audit, or when a food safety alert comes in about a specific parcel — somebody sits down and starts collecting fragments: a call to the carrier, a spreadsheet tab from the cellar supervisor, a photo of the notebook. The traceability exists, but it has to be rebuilt by hand every time. And every rebuild is an hour in which the plant is not producing.

How it fits the IRIS system

iLEAN does not add another database — it seals the cracks between the ones you already have.

Table olive traceability does not break for lack of systems: it breaks because the three worlds do not talk to each other at second zero. iLEAN is the putty that fills the joints between the dock, the cellar and the line, without asking you to change any of the three. You do not knock down the wall, you fill the gap.

Edge reads the container and the code. Connect captures the delivery note and the tank record wherever they live. The agent closes the per-batch dossier and leaves it on the quality manager's desk. The audit starts with everything already on the table.

The three iLEAN pieces applied to preserved table olives:

• Edge — terminals with machine vision (CNN) at size grading and packing. They read size grade, container format and the inline printed batch code without the operator writing anything down. If the code does not match the production order, product is held before palletizing. It works with no network: if the plant loses WiFi, it keeps reading and keeps holding.
• Connect — captures the carrier's delivery note by photo or voice on the dock, reads the tank record from the cellar supervisor's notebook (photographed, dictated or integrated), and also picks up what arrives from outside (an importer's email with a new retailer rule, a WhatsApp from the brine supplier). Everything comes in and becomes usable data.
• Agent — keeps the balance of each tank, cross-references it with the packing line and closes the per-batch file. For a food safety alert on a parcel, it tells you within minutes which packed batches contain it. For an IFS/BRC/FSMA audit, it hands over the dossier already built. It exports nothing on its own — the person validates and signs.

See the full IRIS architecture →

Before and after

Reconstructed traceability vs. continuous traceability with iLEAN

Aspect	Today — reconstructed traceability	With iLEAN Edge + Connect + Agent
Receiving on the dock	Paper delivery note, operator's notebook	Delivery note captured by photo/voice in Connect
Fermentation tank balance	Cellar supervisor's notebook	Continuous, automatic balance
Batch changeover at packing	Paper checklist, shift lead's signature	Edge verifies format + code + recipe
Batch dossier for the retailer	Hours collecting spreadsheets, photos and phone calls	Signed PDF in minutes
Food safety alert on a parcel	A day lost hunting for where it went	List of batches and shipments in minutes
IFS/BRC/FSMA audit	The week before spent firefighting	A living file, ready whenever they ask

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.

• Mid-sized table olive packer, 15-25 fermentation tanks, 2 glass packing lines + 1 pouch line, exporting to the EU and the US.
• Pilot: Connect on the dock and in the cellar + Edge on one packing line + dossier agent. First value expected within a few weeks — the first complete batch dossier generated on its own.
• Indicative payback between 4 and 9 months, depending on audit frequency, retailer demands and the documented cost of each manual traceability rebuild.
• Hard measurable lever: ≥ 30% reduction in quality hours spent rebuilding files, and a response to a parcel-level food safety alert in minutes instead of a day.

And the quality manager's reasonable doubt

“What if the agent gets a dossier wrong?” — hallucination is a problem of free generation, not of anchored tasks. Here the AI invents nothing: it reads a delivery note, cross-references it with the tank record, cross-references that with the container code and packages what already exists. In tasks anchored to a source, the best models brought error below 1.5%^[1]. And even so, no dossier goes out signed without the person. The three safety rings exist precisely for this.

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

Related: olive oil mill traceability · FSMA export to the US · continuous IFS/BRC audit