Non-conformities come from the gaps, not from the process

The problem

Non-conformities come from the paper that cannot be found, not from the process.

A precooked food plant with exports and retail faces several audits a year, each with its own format and vocabulary. All the evidence exists, but every piece is somewhere different: a folder, a spreadsheet, a filing cabinet, somebody's inbox. And here is what hurts most about that work: non-conformities rarely come from the process being wrong. They come from not finding the paper that proves it was right. The auditor asks for the evidence, the plant produces folders, and the documentary gap becomes a finding. On top of that, every traceability query from an export customer opens the same process in miniature: days of searching to answer one specific question.

• A precooked food plant with exports and retail faces several audits a year — certification, retail distributor, export customer — each with its own format and vocabulary.
• All the evidence exists, but every piece is somewhere different: cleaning records in a folder, tunnel temperatures on a form, laboratory controls in a spreadsheet, delivery notes in a filing cabinet, a supplier certificate in somebody's inbox.
• Non-conformities rarely come from the process being wrong. They come from not finding the paper that proves it was right. The auditor asks, the plant produces folders, and the documentary gap becomes a finding.
• And every traceability query from an export customer opens the same process in miniature: days of searching to answer one specific question about one batch.

How it fits the IRIS system

Agents on top of what Connect and Edge already capture — the evidence organizes itself.

Agents on top of the evidence Connect and Edge already capture.

Every critical changeover and every batch aggregate their own dossier the moment they happen: changeover images, tunnel curve, laboratory controls, goods-in temperatures, signatures with timestamps. Anything rebuilt afterwards inherits the gaps of whatever nobody wrote down, and it is precisely in the rebuilding that findings appear. The plant does not produce more evidence than today; it stops losing the evidence it already produces.

• Every critical changeover and every batch automatically aggregates its before/after photos of critical points (case 10), the tunnel curve (case 2), laboratory controls (case 6), goods-in temperatures (case 7) and signatures with their timestamps.
• Agents maps each piece of evidence against each standard's requirements and composes the dossier for a period or a batch on demand.
• It spots what is missing before the auditor does, while there is still time to resolve it.
• The auditor logs in and downloads without anyone on the floor collecting anything.

See the full IRIS architecture →

Before and after

Today's preparation versus the permanent dossier

Aspect	Today	With iLEAN Agents
Audit preparation	Weeks in quality	Virtually zero
For each auditing customer	Repeated in full, in their format	Same evidence, another view
Where the evidence is	Folder, spreadsheet, cabinet, inbox	One dossier per batch
A documentary gap	Found by the auditor	Spotted while there is time to resolve it
A traceability query from an export customer	Days	Minutes
The auditor's access	Folders on a table	Logs in and downloads

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, depending on the number of audits per year and the current preparation load.
• Continuous protection of the certifications that retail and export contracts depend on, which is the value that never appears in a time study.
• From weeks of preparation per audit to virtually zero, and the same evidence serving the certification body, the retail distributor and the export customer.
• And a traceability query answered in minutes, which is what an export customer remembers when the next contract comes up.

estimated payback of 6 to 12 months depending on the number of audits per year and the current preparation load, plus continuous protection of the certifications that retail and export contracts depend on. *Estimate to be validated*.

And the fair question from the production manager

“Isn't this just another document manager?” — the difference is direction. A document manager stores what somebody decides to upload, after the fact, and inherits the same gaps. Here the dossier closes the instant the changeover or the batch happens, from evidence captured at the line. Mapping each piece of evidence against a standard's requirement list is an anchored task, where the best models drop below 1.5% error [1], and the quality manager reviews what the system flags as missing before the auditor ever asks.

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