Paperless lot startup

The problem

The two papers that open a lot are the last ones to reach a screen.

The process sheet defines whether the mold started with approved (PPAP) parameters; the traveler card is the only physical proof of the lot as it crosses injection, assembly and packaging. Today both can go hours without being digitized. If there's a deviation, it surfaces too late — when the auditor asks for traceability or the customer reports a defect, reconstructing what happened means searching scattered paper.

• The process sheet says whether that mold started with the parameters approved in PPAP: melt temperature, injection pressure, holding time, cooling time, screw position.
• The traveler card is the only physical proof of the lot as it crosses injection, assembly and packaging, and it travels with the tote, not with the system. If it gets wet, torn or left behind, the lot loses its identity.
• Between the press starting and somebody typing those values, hours go by, and over a weekend considerably more.
• During that window the lot is physically running and digitally does not exist, which is exactly when a customer complaint or an auditor asks about it.

How it fits the IRIS system

Connect in photo mode — two photos at startup, nothing else changes.

Connect photo mode. The operator takes two phone photos at startup. A grounded language model extracts mold, cycle parameters, lot, shift and operator, and saves them to iLEAN's central memory in zero seconds. Zero habit change: the operator keeps filling out the same paper they already know.

This is the shortest of the twelve to deploy and the one that meets least resistance on the floor, because it asks nothing new of the operator: the same paper they already fill in, plus a photo. A grounded model pulls out mold, cycle parameters, lot, shift and operator, and the lot is open in the central system before the press finishes its first full cycle. That is usually why it is the first one signed off.

See the full IRIS architecture →

Before and after

Lot startup on paper versus lot startup captured

Aspect	Today	With iLEAN Connect
When the lot exists digitally	Hours after the first shot	Second zero
Check against approved PPAP parameters	Nobody does it at startup	Automatic, at startup
Traveler card	A card in a tote	A card and its digital twin
Finding a lot from three months ago	Search the paper archive	A query from any screen
Deviation written by hand at startup	Lost unless typed	Extracted and tied to the lot
What the operator does differently	—	One photo of each document

hours of latency until someone transcribes → zero seconds. Lot found by searching a physical card → instant web query from any screen.

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, depending on how many lot startups a day the plant runs.
• Transcription time given back to the line supervisor, who is the one doing it today.
• From hours of latency — more over a weekend — to zero seconds.
• A lot found by query instead of by hunting for a card that may already have left with the tote.

estimated payback 4-9 months depending on daily startup volume, recovered transcription time for the line supervisor. Estimate to validate.

And the fair question from the production manager

"What if it reads a cooling time wrong off the sheet?" — pulling fields out of a document with a fixed layout is an anchored task, where the best models drop below 1.5% error [1]. And there is a second net: the values land on the cell tablet as a summary and the supervisor confirms or corrects them before anything crosses into the central system.

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