Resin drying straight to ERP

The problem

Drying decides the part, and it is recorded where nothing can see it.

Two parallel worlds — the real spreadsheet, where the live drying data lives, and the official ERP, missing that data or receiving it late, by hand, from someone dedicated to copy-pasting.

• Hygroscopic resin that reaches the screw with residual moisture above spec comes out of the mold with surface streaks, bubbles and mechanical properties below what the drawing promises the customer.
• So the plant controls it properly: desiccant dryer, dew point, hopper temperature and hours per batch, all logged batch by batch by the process manager — in a spreadsheet.
• The ERP never sees that spreadsheet. It gets copied across late, by hand, by somebody whose day includes copy-pasting cells.
• The result is a broken chain exactly where it matters most: the part that failed cannot be tied back to the drying batch that produced it.

How it fits the IRIS system

Connect listening inward — it watches the folder, not the person.

Connect "listen inward" mode — watches the shared folder (not the person), reads the delta when the file is saved, parses it with a grounded model, validates on screen, posts to the ERP via API.

Nobody is asked to abandon their spreadsheet, which is why this case survives contact with the floor. The file keeps being the working tool; iLEAN reads the delta each time it is saved, parses it with a grounded model, shows it for validation and posts it to the ERP through its own interface. The ERP stops being a place data arrives at two days late.

See the full IRIS architecture →

Before and after

The drying record today versus the connected record

Aspect	Today	With iLEAN Connect
The drying log	A spreadsheet the ERP never sees	Read cell by cell as it is saved
Dew point and residual moisture	Written down, then filed	Tied to the batch that was molded
Dryer-to-part chain	Broken	Complete, dryer by dryer
A part that failed for moisture	Nothing to look back at	The drying curve of that exact batch
Who needs an ERP seat	Everyone who logs drying data	Only the manager who validates
Quality and process licenses	One per person entering data	50-80% fewer

drying data disconnected from the production lot → full drying-moisture-to-part traceability. Manual transcription → zero.

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, before counting the license saving.
• A typical reduction of 50-80% in quality and process seats, because only the validating manager needs to log in.
• Manual transcription of the drying record down to zero, and with it the errors it introduces.
• Full traceability from dew point and drying hours to the part number that was molded, which is what a customer asks for when a part fails on properties.

— *CFO angle*: if several process and quality people log into the ERP today just to enter drying data, that's several licenses; with Connect, only the validating manager needs one. Typical reduction of 50-80% in quality/process module licenses. Estimated payback 4-9 months. Estimate to validate.

And the fair question from the production manager

"Our spreadsheet is a mess of merged cells and notes in the margin — can it read that?" — it can, because reading a sheet with a stable layout is an anchored task, where the best models drop below 1.5% error [1]. And the delta is not posted blind: it goes to the tablet as a summary and the process manager confirms it, so a note in the margin becomes a flag rather than a bad number in the ERP.

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