The metrology lab spreadsheet, straight to the ERP

The problem

The lab measures in minutes and the ERP learns about it weeks later.

- Lab data is critical for any PPAP or audit, but the ERP still requires manual transcription, costing half a day a month and leaving weeks of lag between real and ERP data. - Every lab technician who logs into the ERP just to transcribe needs their own seat.

• The metrology lab keeps a spreadsheet with electrical test, torque and calibration results: winding resistance, insulation resistance, alternator output curves, regulator setpoints, gauge calibration dates.
• That data is critical for any PPAP or audit, but the ERP still requires manual transcription, costing half a day a month and leaving weeks of lag between the real value and the one in the system.
• During those weeks, a lot can be released against an ERP that does not yet show the lab result that would have held it.
• And every lab technician who logs into the ERP only to transcribe needs their own seat in the quality module.

How it fits the IRIS system

Connect listening inward — it reads the file as soon as it is saved.

Connect 'listening inward': the system reads the shared file as soon as it's saved, parses it with a grounded LLM, and after human validation posts it to the ERP automatically via API.

The lab keeps its spreadsheet, its columns and its way of working. Connect reads the saved file, not the person, and the ERP receives the validated result minutes later instead of at the end of the month. Release decisions then rest on the lab's latest result, not last month's.

See the full IRIS architecture →

Before and after

The lab spreadsheet today versus the spreadsheet read by Connect

Aspect	Today	With iLEAN Connect
Winding and insulation resistance results	In the spreadsheet only	Posted to the ERP against the lot
A failed hipot sample	Invisible to release until transcribed	Visible in the ERP before the lot ships
Gauge calibration status	Updated when someone remembers	Current as soon as the file is saved
PPAP electrical and dimensional data	Copied and reconciled by hand	Exported from the ERP as recorded
Who needs an ERP login	Every technician who transcribes	Only the person who validates
Spreadsheet layout	Owned by the lab	Still owned by the lab, read as it is

Half a day a month of transcription and weeks of lag → data in the ERP within minutes of validation.

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 4-8 months, driven mainly by license savings.
• Direct savings of -50 to -80% on ERP seats in the quality/lab module, because technicians no longer log in just to transcribe.
• Half a day a month of transcription and weeks of lag become data in the ERP within minutes of validation.
• We validate the seat count and the savings with your IT team before committing, on your real license structure and on the technicians who today log in only to type results.

Direct savings of -50 to -80% on ERP seats in the quality/lab module, with estimated payback 4-8 months. Estimate to be validated with IT.

And the fair question from the production manager

“What if it posts a wrong resistance value to the ERP?” — reading a spreadsheet whose columns and units are known is an anchored task, where the best models stay below 1.5% [1] error. Nothing is posted until a person validates it, and a value outside the physical range for that test, such as a negative insulation resistance or a torque above the tool's capacity, is stopped before it even reaches validation.

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