The audit dossier, always ready

The problem

Before each audit, someone rebuilds months of changes by hand.

Quality audits require evidence that today gets rebuilt manually, risking nonconformities from documentation gaps.

• Format changes, color changes, kiln curve adjustments and batch reclassifications all leave a trace — on paper, in spreadsheets, in the ERP and in people's memory. None of those places was designed to answer an auditor. And none of them stores a photo of what the line looked like before and after the change.
• The day the quality auditor arrives, someone spends weeks putting that trace back together for the period under review, usually the same people who should be running quality on the floor.
• Whatever was not written down at the time cannot be rebuilt. That is where nonconformities from documentation gaps appear. And they appear in front of the auditor, not before.
• And a tile plant usually answers to more than one scheme: the quality management system, product certification and marking, and sometimes environmental declarations. Each one asks for the same evidence in a different order. The work is repeated for each of them. And each visit starts the reconstruction again from scratch.

How it fits the IRIS system

Edge captures photo and data per change; Agents organizes the dossier by standard.

Edge captures photo + data per change; Agents organizes it by certification standard and generates the dossier structured by period.

Every critical change closes its own record at the moment it happens, with before and after images and the signature of whoever approved it. The audit stops being a reconstruction and becomes a filtered view of what already exists. The same dossier answers the certification body, the product certification and a customer audit. What used to take weeks becomes a review. And the quality team spends that time on the floor, where it belongs.

See the full IRIS architecture →

Before and after

Certification audit preparation today versus the dossier built per change

Aspect	Today	With iLEAN Edge + Agents
Weeks spent before the auditor arrives	Weeks	Zero
Before and after photos per change	Rarely taken	Captured automatically
Format and color change records	Spread across paper and files	One dossier per change
Several certification schemes	Prepared one by one	Same evidence, indexed per standard
Documentation nonconformity risk	Found during the audit	Eliminated at the source
Period dossier	Assembled by hand	Generated on demand

Weeks of pre-audit reconstruction → zero. Documentation-nonconformity risk → eliminated.

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 how often you are audited. The more schemes and visits, the shorter it gets.
• Continuous protection of the plant's certifications, which are what keep it on distributors' and specifiers' lists. Losing one is far more expensive than any preparation time.
• Weeks of pre-audit reconstruction drop to zero. The quality team stops dreading the date in the calendar.
• And the risk of a nonconformity from a documentation gap is eliminated at the source, not discovered on the auditor's day. A missing record is flagged when it happens, while the people involved still remember what was done.

Continuous protection of certifications, estimated payback 6-12 months depending on audit frequency. *Estimate to validate*.

And the fair question from the production manager

“Isn't this just another document management system?” — the difference is the direction. A document manager stores what someone decides to upload afterwards, with the same gaps. Here the record is created at the moment of the change, and sorting it by standard is an anchored task, where the best models drop below 1.5% error [1]. The quality lead reviews the dossier before it is handed over. Nothing is delivered to an auditor without that review.

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