The laboratory releases batches without anyone retyping into the ERP

The problem

The lab releases on a spreadsheet; the ERP finds out days later.

The lab works in its sheet because it is fast and flexible; the ERP demands structured data. In between sits a person transcribing: in the order of half a day a month of work that adds nothing. The lag has commercial consequences. Sales commits deliveries on batches the ERP does not yet treat as released, even though the lab approved them that morning. Invoicing slips because of a transcription problem, not a quality one. And there is a third cost, the one visible on the ERP invoice: for lab data to reach the system today, you need a quality-module user for every analyst who types.

• The analysts run density, color, water content, flash point, viscosity, purity and pH, and type the results into a sheet that is fast, flexible and, in practice, the batch release system.
• The COA comes out of that sheet, and from there someone retypes it into the ERP quality module. Half a day a month of an analyst's time that adds nothing, and a transcription error waiting to reach a customer.
• Sales commits deliveries on batches the ERP does not yet treat as released, although the lab approved them that morning. Invoicing slips for a transcription reason, not a quality one.
• For lab data to reach the ERP today, every analyst who types needs a quality-module user. That line on the ERP invoice is the one the finance director sees first.

How it fits the IRIS system

Connect in listening-inward mode — the file, not the person.

Connect in "listening inward" mode.

The lab keeps its sheet, in the format it has today. iLEAN reads it when it is saved, compares each parameter with the product specification, drafts the COA and waits for the quality manager to validate. One license validates instead of five typing, and the batch is released in the ERP the same morning the lab released it. The finance director sees it on the next ERP invoice; sales sees it the same week, in batches that stop being committed before the ERP treats them as released.

• iLEAN observes the shared folder and reads the file when it is saved. It observes the file, not the person.
• A model grounded in the sheet layout extracts the parameters: density, color, water content, flash point, viscosity, purity, pH.
• It checks them against the product specification and flags any out-of-range value.
• It drafts the batch certificate of analysis.
• The quality manager validates or corrects, and only then does the data cross into the ERP quality module through its interface.

See the full IRIS architecture →

Before and after

Today's laboratory versus the connected laboratory

Aspect	Today	With iLEAN Connect
Lab → ERP lag	Days	Minutes
Retyping the COA	Half a day a month	Zero
Out-of-spec value	Caught at review, if at all	Flagged as it is entered
Certificate of analysis	Rebuilt by hand per batch	Drafted, validated, posted
Invoicing	Waits for the transcription	Follows the release
Quality-module licenses	One per analyst who types	One: the manager who validates

from half a day a month of transcription to zero · from days of lag between lab and ERP to minutes · from certificates rebuilt by hand to certificates generated and validated · from out-of-range values caught at review to values flagged as they are entered.

Impact estimate

Impact estimate — to be validated against last year's ERP billing.

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 3-7 months, with the license line as the anchor.
• 💰 CFO angle: the ERP bills per named user. If five analysts log in purely to type results, that is five quality-module licenses: typical ratio 5 down to 1, 50-80 % less recurring cost in that module.
• Half a day a month of analyst time released, and a transcription error that no longer reaches a customer's COA, which is the kind of error that costs a credential, not a correction.
• Invoicing brought forward by earlier release: the batch the lab approved at nine is released in the ERP at nine, not on Thursday.

recurring saving on quality-module licenses — typical ratio of 5 licenses down to 1, between 50% and 80% less recurring cost in that module — plus the transcription time released and the invoicing brought forward by earlier release. Estimated payback of 3 to 7 months. *Estimate to validate against last year's ERP billing.*

And the fair question from the production manager

“That sheet has been changed by five analysts over ten years — does it hold up?” — that is the anticipated case: parsing does not rely on fixed cell positions, it interprets columns, units, merged cells and notes without asking anyone to standardize anything. Extracting parameters from a sheet with a known structure is an anchored task, where the best models drop below 1.5% error [1], and the quality manager validates every COA before it crosses into the ERP.

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