Two worlds that finally talk: spreadsheet and ERP

The problem

The control plan lives in a spreadsheet the ERP has never seen.

two parallel worlds — the real spreadsheet (live data required by IATF 16949) and the official ERP (data transcribed weeks late by someone spending half a day a month copy-pasting).

• Every batch, the metrology lead records what the control plan demands: outside diameter, wall thickness, length, ovality, straightness, plus weld checks such as flattening or expansion tests.
• That spreadsheet sits in a shared folder. It is the live record, the one the metrology lead trusts and the one IATF 16949 cares about. Several people have edited it for years, and it has grown tabs and notes along the way.
• The ERP holds a transcribed version that arrives weeks late, typed by someone who spends half a day a month copying and pasting.
• When an OEM asks for capability on a characteristic, or a PPAP needs dimensional results, the two worlds have to be reconciled by hand, against the clock, and every transcription error surfaces at exactly that moment.

How it fits the IRIS system

Connect listening inward — it reads the shared folder, not the person.

Connect's "listen inward" mode watches the shared folder, not the person, absorbing the data in real time with a grounded LLM; after validation, it flows to the ERP via API.

The metrology lead keeps the spreadsheet they trust. iLEAN absorbs each cell as it is saved and, after validation, it flows to the ERP via API — so the official record and the real one stop being two different things. The capability an OEM asks for at each review is then a query, not an afternoon's work. And the data your control plan demands is where the auditor expects to find it.

See the full IRIS architecture →

Before and after

Two parallel records versus one record that flows

Aspect	Today	With iLEAN Connect
Lag between measurement and ERP	Weeks	Zero latency
Monthly copy-paste	Half a day	Zero
Flattening and expansion test results	In a tab nobody else opens	Tied to the batch in the ERP
Capability on outside diameter	Recomputed for each request	Current on the day it is asked
A mistyped wall thickness	Surfaces at a PPAP	Never typed
Seats in the quality/lab module	One per person who types	Reduced by 50-80%

weeks of lag → zero latency. Half a day a month transcribing → zero.

Impact estimate

Estimated impact — to validate against your actual license invoice.

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.

• 💰 CFO angle: the ERP bills per named user. Everyone who logs into the quality or lab module purely to type results is a seat: a typical 50-80% reduction, recurring every year.
• Estimated payback 4-9 months, before counting the license saving.
• From weeks of lag to zero latency, and half a day a month of transcription down to zero.
• Dimensional results ready for a PPAP or an OEM review without reconciling two records, which is where most of the time goes today.

typical 50-80% reduction in quality/lab-module licenses, estimated payback 4-9 months. *Estimate to validate*.

And the fair question from the production manager

“Our metrology sheet has a tab per program and notes in every margin — will it cope?” — yes, because parsing does not rely on fixed cell positions: it reads headers, units and tabs as they are. Mapping a spreadsheet with a stable structure is an anchored task, where the best models drop below 1.5% error [1], and the metrology lead validates the batch before anything is pushed. Nobody is asked to standardize the sheet first.

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