Visual evidence before signing off clean

The problem

The cleaning sign-off is the weakest link of a shared line.

Leftover ink or varnish from the chemical run can contaminate the first food-contact batch. A cross-contamination event puts the BRC Packaging Materials certification at risk overnight.

• A line producing both chemical and food-contact packaging requires certified cleaning before every switch to a food-contact run.
• Today it depends on a human sign-off with no evidence: a ticked box on a sheet, under pressure because the changeover sits on the shift's critical path.
• Leftover ink or varnish from the chemical run can contaminate the first sheets of the food-contact batch. The residue is not always visible to a person in a hurry, and it does not announce itself until much later.
• A single cross-contamination event puts the BRCGS Packaging certification at risk overnight, and with it the food-contact contracts.

How it fits the IRIS system

Edge with JIDOKA AI — the cameras check, the quality lead signs on evidence.

Edge + JIDOKA AI. Fixed cameras on the inker and varnish unit compare the current state against the clean reference state. The line won't start until every camera reports OK and the quality lead signs off based on visual evidence.

The camera does not replace the sign-off; it gives it something to stand on. The quality lead still decides, but on images of each critical point instead of on trust. That is also what lets the signature stand up later in front of an auditor.

See the full IRIS architecture →

Before and after

The chemical-to-food changeover signed blind versus signed on evidence

Aspect	Today	With iLEAN Edge
Cleaning sign-off	A tick on a checklist	Each critical point compared with its clean reference
Inker and varnish unit	Checked by eye	Fixed cameras on both
Leftover chemical ink or varnish	Found in the first food sheets, if at all	Detected before startup
Line start	When someone signs	Only when every camera reports OK
Proof for the auditor	A signature	Before and after images per changeover
Certification exposure	Overnight, on one event	Each switch backed by evidence

Blind sign-off with cross-contamination risk → sign-off based on visual evidence at each critical point.

Impact estimate

Impact estimate — the payback depends on your food-contact mix.

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.

• We do not fix a payback range here: it depends on the weight of the food-contact line in your mix, and we size it with your numbers.
• What it protects is the certification behind hundreds of thousands of euros in food-contact contracts. Losing it is not a cost per run, it is the loss of the food-contact business at that site.
• Blind sign-off becomes sign-off on visual evidence at every critical point of the changeover. The quality lead signs knowing what they are signing, and can prove it later.
• And the evidence of each switch is stored, which is exactly what the next audit asks for.

Protects a certification worth hundreds of thousands of euros in food-contact contracts; payback depends on the weight of the food-contact line in the mix. Estimate to validate.

And the fair question from the production manager

“What if the camera blocks the line for nothing?” — comparing the inker and the varnish unit against their own clean reference state is an anchored task where the best models drop below 1.5% error [1]. In case of doubt the system does not block silently: it escalates to the quality lead, who signs on the images. It stops on a visible discrepancy, not on model uncertainty. And because the check takes seconds, it does not lengthen a changeover that already sits on the critical path.

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