You already own the good pieces — they just needed stitching

The problem

The classic “almost integrated”: three ends joined by a naming convention.

This is the classic "almost integrated". The CMM exports its report to a folder, the programmer saves the CAM file in another, and the job order lives in the ERP or on paper. Nothing joins them except a naming convention, and conventions break: a rush job, a misspelled name, a revision that changes halfway. When months later you have to prove which program and which drawing a part was made against, the answer depends on one person's memory — and that person is not always around.

• The CMM exports its report to a folder, the programmer saves the CAM to another, and the work order lives in the ERP or on paper.
• Nothing joins them but a naming convention, and conventions break: a rush job, a misspelled name, a revision that changes halfway.
• When months later you have to prove which program and which drawing a part was made against, the answer depends on one person's memory.
• And that person is not always there.

How it fits the IRIS system

Connect acts as the stitching between the three ends — nothing to buy.

Connect acts as the seam, by API or export, between the three ends: CMM report ↔ CAM program ↔ job order and revision.

Here the hardware is already bought and depreciated. The CMM issues a correct report and CAM stores the program: all that was missing was for them to know which order what they just produced belongs to.

• When a measurement closes it is bound to the job, to the current revision and to the specific program that produced it.
• What happens in one is reflected in the others automatically.
• Without replacing the CMM, changing the CAM or touching the ERP.

See the full IRIS architecture →

Before and after

Naming convention vs. automatic link

Aspect	Today	With iLEAN Connect
Report ↔ program ↔ order link	Manual naming convention	Automatic and auditable
When the convention breaks	Nobody notices	There is no convention to break
Current revision	Inferred	Tied to the measurement
Historical reconstruction	Depends on one person	Direct lookup
CMM software	—	Untouched
Equipment investment	Already made	None new

report↔program↔job link held by a manual naming convention → automatic and auditable link. Historical reconstruction depending on one person → direct query.

Impact estimate

Impact estimate — to validate against 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.

• Shops that already run a CMM and CAM but disconnected from the work order.
• Indicative payback between 5 and 10 months, with low CAPEX because the hardware is already bought and depreciated.
• To confirm first: that the CMM software exposes automatic export or an API. It is the only technical unknown in the case.
• The real return shows up in the audit, when you have to prove which program and which drawing a specific part was made against.

estimated payback 5-10 months, with low CAPEX because the hardware is already bought and depreciated. *Estimate to be validated* once the CMM software is confirmed to expose automatic export or an API.

And the fair question from the production manager

“What if our CMM is old and exposes nothing?” — that is the first check and it happens before anything is committed. Most measuring software allows automatic export to a folder, which is enough. If a particular unit does not, that gets said and another route is used — normally the camera over the panel, which is another case in this matrix — rather than forcing the integration.

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