The lab spreadsheet poured into the ERP, without retyping

In a long steel mill the metallurgist fills in a spreadsheet with the heat chemistry and the mechanical tests. That spreadsheet is where the real data lives — but the ERP QM module demands transcription with hours of lag. iLEAN Connect reads it as it is written, the metallurgist validates it and it is poured into the ERP QM module over API. Heat release goes from hours to minutes.

‹ See all 12 long steel cases

Metallurgist in the lab of a long steel mill saving the heat chemistry and mechanical test spreadsheet into a shared folder, with a screen alongside showing the validation before the transfer into the ERP QM module
The problem

Two parallel worlds: the real lab spreadsheet and the official QM module.

In a long steel mill, the lab keeps its own spreadsheet: the heat chemistry — 14-18 elements — and the mechanical tests, analysis by analysis. It is the data the metallurgist actually works with, updated to the minute in the shared folder. The ERP QM module, by contrast, receives those same results typed in by hand by somebody copying from the spreadsheet into the system — with hours of lag.

  • Two parallel worlds — the lab spreadsheet, where the live data is, and the ERP QM module, where somebody pecks the figures in with hours of lag. The system the corporation treats as "the truth" runs behind the reality of the lab on every heat.
  • Heats waiting for release with the certificate already issued — the analysis is done and the certificate exists, but the heat is not released until somebody moves the data into the ERP. In the lab VSM, that stretch is pure waste: release lead time inflated by transcription, not by testing.
  • Half a working day a month of an engineer sunk into transcription — not analyzing, not improving, not releasing heats faster: typing what the spectrometer and the tensile machine had already measured hours earlier.
  • An IATF auditor with uncross-referenced data — when the audit asks to trace an analysis from the original sheet to the official record and the dates do not line up, the usual defense — "it is in the spreadsheet, it had not been transferred yet" — is exactly what the auditor does not want to hear.

The cost is not only audit or lead time: it is licenses. For that typing to be possible, several people need access to the ERP QM and PP modules — not because they manage quality or production in the system, but because it is the only way to get their own analysis data in.

How it fits the IRIS system

Connect listens inbound: it watches the folder, not the human.

The lab spreadsheet does not need to disappear or change format. It needs the figure that already exists on that sheet to reach the ERP QM module without anyone retyping it, and without the heat waiting for release with the certificate already issued. That is what Connect is for.

Connect watches the shared folder where the lab saves its spreadsheet — it does not watch the metallurgist filling it in. When the file is saved, it reads the delta, presents it to the metallurgist on screen and, after validation, pours it into the ERP QM module over API. Heat release becomes an event at zero latency.

How Connect works on the lab spreadsheet of a long steel mill:

  • Connect watches the folder, not the human — it points at the network location where the lab saves each heat's spreadsheet. It does not interfere while it is being filled in and it times nobody: it simply detects when the file has been saved.
  • It reads the delta on save — as soon as the spreadsheet is saved, Connect identifies which rows are new since the last transfer and extracts them, without reprocessing the entire history each time.
  • An LLM anchored to the real template — the model generates no free data: it parses exactly what is in each cell — the 14-18 elements of the chemistry and the mechanical tests — against the lab's real template, including its own columns and its usual ordering.
  • A delta summary to the metallurgist — before anything reaches the ERP, the metallurgist sees a summary of what was read on screen: they confirm or correct. The data does not enter on its own.
  • Validation and transfer over API — once validated, Connect writes the result directly into the ERP QM module, with full traceability of who validated what and when. Heat release goes from hours to minutes, and in the lab VSM the release lead time falls by 30-60% (estimate to be validated).

See the full IRIS architecture →

Before and after

Spreadsheet and QM module disconnected vs. spreadsheet and QM module with Connect in between

AspectSpreadsheet and QM module disconnectedWith iLEAN Connect
Spreadsheet-to-ERP lagHours per heatZero latency, on save
Manual transcriptionHalf a working day a month of an engineer copying and pastingZero — automatic transfer over API after validation
IATF auditor cross-referencing spreadsheet and ERPData never cross-referencedAutomatic cross-reference, heat by heat
Release lead time (lab VSM)Hours with the certificate already issuedMinutes — a 30-60% reduction (estimate to be validated)
QM and PP module licenses5-6 users who only key in data1 license — the metallurgist who validates
Heat-quality-certificate traceabilityManual reconstruction when requestedRecorded on every transfer, queryable instantly
Impact estimate

Impact estimate for your plant — to be validated with your own 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.

  • Long steel mill with its own lab, heat analyses in a spreadsheet (chemistry and mechanical tests) and a corporate ERP whose QM and PP modules are billed per active user.
  • If 5-6 people access the QM and PP modules today only to key in analysis data, those are expensive licenses paying for typing alone. With Connect only the license of the person who validates — the metallurgist — is needed: from 5-6 licenses to 1.
  • Typical reduction of 50-80% in ERP QM and PP module licenses — not a one-off project saving but a recurring annual saving defensible in committee, a cost that disappears from the invoice. Estimate to be validated against your real ERP invoice.
  • Estimated payback of 4 to 9 months, counting the deployment of Connect on the lab folder and the anchoring to the real heat template. Estimate to be validated against your numbers.
  • Side lever: with the analysis live per heat in the QM module from second zero, release lead time falls from hours to minutes and the heat-by-heat cross-reference the IATF auditor asks for stops being a manual reconstruction and becomes a query. That relief of quality workload is not included in the license calculation.

And the fair question from the finance director

"What if the LLM misreads a spreadsheet cell and a wrong carbon percentage reaches the QM module?" — hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI merely recontextualizes a specific figure from one medium to another (reading an Excel cell and moving it into the corresponding field of the ERP QM module), the best models brought the error below 1.5% [1]. And even then, the critical call is not made alone: Connect holds the transfer until the metallurgist validates on screen. The three safety rings exist precisely for this.

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

Frequently asked questions

What people ask about moving the lab spreadsheet into the QM module with Connect

How does Connect watch the lab folder without changing the metallurgist's work?

Connect watches the shared folder where the lab already saves the heat spreadsheet — it does not watch the metallurgist filling it in. When the file is saved at the end of the analysis (chemistry across 14-18 elements and mechanical tests), Connect detects the delta — only the rows new since the last transfer — and reads it with an LLM anchored to the lab's real template. The metallurgist keeps working with the same spreadsheet as always, the same tabs and the same per-heat habit; there is no new form to learn and no screen to install next to the spectrometer or the tensile machine.

What happens if the lab changes the heat spreadsheet template?

If the lab adds a column, renames an element of the chemistry or reorders the tabs by steel grade, Connect does not break silently: the LLM anchored to the template detects the variation and, if the change is ambiguous, flags it for the metallurgist to review rather than transferring a misinterpreted figure into the ERP QM module. A template adjustment requires no new integration and no ticket to IT — the reading is reconfigured, not the plumbing. And for quality, the change is recorded: which template was active in each transfer and from which heat it applies — traceability an IATF auditor can cross-reference.

How is the real saving on QM and PP module licenses calculated?

The logic is about license consumption, not hours: the ERP QM and PP modules are billed per active user, and if 5-6 people access them today only to key in analysis data, those are expensive licenses contributing no quality management — only typing. With Connect it is enough for one single person — the metallurgist who validates before the transfer — to keep the license. That concentration typically means a 50-80% reduction in QM and PP module licenses: a recurring annual saving defensible in committee, with an estimated payback of 4-9 months. It is an estimate to be validated against your real ERP invoice — which is why the first step is calculating it on your numbers.

What validation exists before the data enters the ERP?

No analysis reaches the ERP QM module without first passing through the metallurgist's screen: they see a delta summary of what Connect read from the spreadsheet — heat, steel grade, the 14-18 elements of the chemistry and the mechanical tests — compare it with the original figure and confirm it before the API transfer fires. Connect does not write into the ERP on its own — it prepares the data and leaves it ready to validate, and the person validates. Every transfer is traced per heat: what was read, who validated it and when — exactly the spreadsheet-to-ERP cross-reference the IATF auditor asks for and does not find today.

Does it work with any ERP or QM module?

Yes. Connect transfers the validated figure over API, so it does not depend on a specific ERP brand or QM module: it works the same with the QM module of a corporate ERP, with a departmental quality system or with a custom development, as long as it exposes an endpoint to write the heat result to. It does not replace your ERP and it does not migrate data — it automates the stretch a person does by hand today: reading the lab spreadsheet and typing it into the official system, with hours of lag while the heat waits for release.

Let's talk

Cut QM and PP module licenses without losing traceability — we will calculate your real license saving.

We work on your plant's real data, not ours. Assessment with no commitment.

Calculate your real license saving ‹ See all 12 long steel cases See steel