The travel card no longer gets lost on paper

The problem

The card that ties the lot to its program travels by hand.

The travel card and the first-off control piece are filled by hand and travel physically through 3-4 stations. If it gets lost or filled in badly, the lot-to-program traceability IATF 16949 requires is lost until someone reconstructs it.

• Every batch opens with a travel card: program, gauge, cut length, destination OEM, shift and operator. It is filled in at the tube mill and then travels physically through 3-4 stations — forming, welding, sizing, cut-to-length.
• The first-off control piece is recorded the same way: outside diameter, wall thickness and the weld check, written on paper next to the line.
• Until someone transcribes it at the end of the shift, the lot exists on the floor but in no system that quality or the OEM can query.
• If the card gets lost, soaked in coolant or filled in wrong, the lot-to-program traceability IATF 16949 requires is broken until someone reconstructs it from memory and rack tags.

How it fits the IRIS system

Connect in photo mode — the card stays on paper, the lot goes digital.

Connect photo mode. The operator captures the card with a phone or industrial tablet. A grounded LLM extracts structured fields (program, gauge, length, destination OEM, shift, operator) and saves them to iLEAN's central memory. Zero habit change for the operator.

It is the simplest of the twelve to deploy and usually the first one approved, because nothing on the floor changes except one photo. The card keeps traveling with the bundle, as it always has; what changes is that the lot it describes already exists in the central system when the first tube leaves the cut-off saw.

See the full IRIS architecture →

Before and after

The travel card on paper versus the travel card captured

Aspect	Today, on paper	With iLEAN Connect
When the lot becomes queryable	At shift end, if someone transcribes it	At startup, zero latency
Lot tied to program and gauge	Only on the physical card	Structured fields in central memory
A card lost between stations	Traceability rebuilt by hand	The photo already exists
First-off control piece	Handwritten, filed by the line	Stored with the lot it released
Destination OEM of a bundle	Read off the tag on the rack	One query by lot number
Operator's routine	—	Same card, plus one photo

traceability that depends on the physical card not getting lost → a digitally queryable lot from zero latency.

Impact estimate

Estimated impact — to validate with your own 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 lots the plant starts per day.
• The saving comes from the production lead's transcription time, which today goes into copying cards at shift end.
• Traceability stops depending on a physical card not getting lost: the lot is digitally queryable from zero latency.
• And a lost or damaged card stops being a traceability gap that has to be explained to an auditor.

estimated payback 4-9 months depending on daily lot volume, recovering the production lead's transcription time. *Estimate to validate*.

And the fair question from the production manager

“What if it misreads the gauge or the program number on a handwritten card?” — reading fields from a card with a known layout is an anchored task, where the best models drop below 1.5% error [1]. And there is a second net: every field is checked against the programs active in the ERP, and the operator confirms the summary on the line tablet before the lot is recorded.

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