Lab Spreadsheet Into ERP

The problem

Two worlds: the lab knows, the ERP does not.

The lab fills in a spreadsheet for every batch with moisture, particle size, color and microbiology, in a shared folder the ERP never sees: two parallel worlds with days of lag.

• For every batch the lab records moisture, water activity, particle size, color, protein or active content and microbiology in a spreadsheet saved in a shared folder. The spreadsheet works well; the problem is that it is an island.
• The ERP never sees that folder. Batch release, the certificate of analysis and stock status run days behind what the lab already knows.
• Someone retypes the key values into the quality module, which is why the plant pays for several quality licenses that exist only to move numbers from one screen to another. Every retyped value is one more chance to swap a lot number.
• When microbiology comes back out of spec after the batch has already been shipped or reused as rework in another blend, the link between result and lot is a manual search through files and emails.

How it fits the IRIS system

Connect listening inward — the lab keeps its spreadsheet, the ERP gets the data.

Connect listens inward: watches the shared folder, reads the delta on save, parses it with a grounded LLM, validates on screen with the quality manager, and syncs to the ERP via API.

The lab does not change tool, and that is precisely why this case gets approved quickly. iLEAN watches the folder, reads only what changed on each save, shows it to the quality manager for validation and pushes it to the ERP. The CFO angle is just as concrete: most of the quality licenses were only there to retype what the lab had already written.

See the full IRIS architecture →

Before and after

QC results today versus results synced on save

Aspect	Today	With iLEAN Connect
Lag from lab result to ERP	Days	Zero seconds after validation
Moisture, water activity and particle size	In a spreadsheet only	On the batch record in the ERP
Microbiology result arriving days later	Searched for by hand	Attached to the lot when saved
Certificate of analysis per blend lot	Built by copy and paste	Generated from synced data
Quality-module ERP licenses	Several, for retyping	One
Lab technicians' workflow	—	Same spreadsheet, same folder

Days of lag → zero seconds. CFO angle: from several lab ERP licenses down to 1 → typical 50-80% reduction in quality-module licenses.

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 4-9 months, mostly from the reduction in quality-module licenses.
• A typical 50-80% reduction in quality-module ERP licenses, from several down to one — an estimate to validate against your actual license invoice.
• From days of lag to zero seconds between the lab saving a result and the ERP knowing it, so release decisions are made on current data and batches stop waiting in quarantine for a number that already exists.
• And batch release and certificates of analysis are built from data that is already in the system, not copied from a spreadsheet at the last minute.

Payback 4-9 months · 50-80% reduction in quality-module ERP licenses — estimate to validate

And the fair question from the production manager

"What if it puts a moisture value on the wrong lot?" — mapping the columns of a spreadsheet the lab has used for years is an anchored task, where the best models drop below 1.5% error [1]. And the delta is never pushed blind: the quality manager sees it on screen and validates it, and values outside each parameter's specification are highlighted before anyone approves them.

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