The customer's spec change, caught instantly

The problem

The revision arrives in an inbox; the horn is cut to the old drawing.

The automotive customer changes the geometry of its part, or the titanium supplier announces a delay. The email lands in the inbox of someone who is in a meeting. By the time someone reads it, the horn is already being machined to the previous revision. Rework, lost material and a delivery date at risk.

• The automotive customer changes the geometry of the part to be welded, which changes the horn contour and the nest. The notice arrives as an email with a PDF, addressed to whoever quoted the machine.
• That person is in a meeting, on a FAT or traveling. Meanwhile the horn workshop keeps machining titanium to the previous revision.
• The same happens with supply: the titanium bar supplier or the piezo ceramics vendor announces a delay in a message nobody in planning sees until the material does not show up at the dock.
• Each late read means scrapped titanium, a re-cut horn and a delivery date at risk on a machine the customer needs for a line launch.

How it fits the IRIS system

Connect in chaotic-sources mode — the notice turns into an action on the right order.

With iLEAN Connect in chaotic-sources mode, a dedicated mailbox or WhatsApp Business number receives those notices. iLEAN extracts the intent, cross-checks the active orders in the ERP and finds which parts are affected. It does not just say an email arrived: it alerts engineering and planning with a suggested action, for example holding the order in machining until the new revision is confirmed. People decide.

The value is not in knowing that an email arrived. It is in knowing, seconds later, which open orders carry a horn for that part, at which station each one is, and what the sensible move is before more titanium is cut. Engineering still decides; it just decides before the chips fly, and planning learns about the change at the same moment, not at the next meeting.

See the full IRIS architecture →

Before and after

A notice in someone's inbox versus a notice crossed with the orders

Aspect	Today	With iLEAN Connect
Customer revision notice	Read hours later by one person	Read on arrival
Which horns are affected	Searched order by order in the ERP	Listed by order and station
Horn already in machining	Finished to the old revision	Hold suggested to engineering
Titanium or ceramics delay	Found when the truck does not come	Re-plan proposed the same day
Who acts on the revision	Whoever happens to read it	Engineering and planning, with context
Trace of the change	Buried in a mail thread	Linked to the order

Before: notices read hours later. After: suggested action in seconds.

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 3-7 months.
• From hours of latency to seconds between the customer's notice and an action on the affected orders.
• Less rework and less high-value material lost: titanium horns and custom nests cut to an obsolete revision, each one worth days of machining.
• Delivery dates defended earlier, because a supply delay becomes a re-plan the same day instead of a surprise at assembly.

Payback 3-7 months · from hours of latency to seconds (estimate to be validated). From hours of latency to seconds, with less rework and less high-value material lost. Estimated payback of three to seven months; an estimate to be validated.

And the fair question from the production manager

“What if it misreads the notice and holds the wrong order?” — extracting intent, part number and revision from a customer email and matching them to open orders is an anchored task, where the best models drop below 1.5% error [1]. And iLEAN does not hold anything on its own: it proposes the hold to engineering with the email and the affected orders in front of them, and a person decides.

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