The die changeover validated before the first stroke
A Tier-2 press changes dies several times a day given the multi-reference nature of the mix. With iLEAN Edge, fixed cameras verify a clean bolster, the correct die mounted and alignment; SMED AI guides the operator step by step; JIDOKA AI blocks the start until the cameras clear it plus the process lead signs with a dimensional first-off. Three safety rings active at every order change.
Every die changeover is a mini SMED that concentrates the PPM risk.
With the reference running steadily, a stamping press is stable. The risk concentrates in the die changeover: a multi-reference Tier-2 press changes order several times a day, and every change resets the bolster, the mounted die, the alignment and the first stroke all at once. That instant repeats often — and today it depends on a human signing "ready" without objective visual evidence point by point.
- If it starts wrong, the first lot comes out of tolerance — a bolster with residual offcuts, a die that is not the one on the order or a drifted alignment turn the first-off into out-of-spec parts before anyone notices, and that first lot goes into the Tier-1 container.
- Releasing the changeover rests on a human signature with no structured visual evidence — the process lead signs a checklist run through in a hurry between two references, with no objective record of the real state of bolster, die and alignment.
- When it fails, nobody knows why it failed — with no per-point evidence, the subsequent 8D analysis is reconstruction from memory: every doubtful start is argued about rather than explained, and the same failure returns at the next order change.
A badly verified die changeover translates into the same thing: startup scrap, net SMED time inflated by readjustments, penalized press OEE and compromised PPM. And since the destination is the Tier-1, an out-of-tolerance start does not stay at home: it becomes a complaint under each OEM's requirements.
Edge clears it point by point — JIDOKA AI does not let the press start until all of them say yes.
The die changeover is not solved with a longer checklist for the process lead: it is solved by giving them objective visual evidence of bolster, die and alignment before they sign the release, and by making the press physically unable to start while that evidence is incomplete. That combination is Edge plus JIDOKA AI, with SMED AI guiding the operator — verification of the setup, not of people.
Edge verifies every point of the changeover — clean bolster, correct die mounted per the order and alignment. SMED AI guides the operator step by step. JIDOKA AI does not release the start until every camera clears it. The process lead signs on the visual evidence and the dimensional first-off, not on a blind checklist.
The iLEAN pieces applied to the die changeover:
- Edge — industrial cameras with CNNs trained to recognize "clean bolster" — no residual offcuts or scrap —, "correct die mounted" per the order and "alignment OK", fixed at the points that change with every die changeover. Each camera learns the reference state of that point for every part number in the mix and compares against it at every change.
- SMED AI — guides the operator step by step through the die changeover, so that external operations are executed while Edge verifies, adding no time to the change. The guidance turns changeover knowledge into a repeatable sequence, not one that depends on who happens to be at the press that day.
- JIDOKA AI — receives each camera's verdict and acts as the start gate: while a single point is not cleared, the press stays held. It is not an alert somebody can ignore under schedule pressure; it is an active block on the start until the evidence is complete.
- The process lead's signature on evidence plus first-off — when every camera clears, the process lead receives the summary with the photos of bolster, die and alignment, confirms the dimensional first-off of the incoming part number and signs the release. The signature remains human and mandatory — what changes is that it is no longer given on a checklist run through in a hurry, it is given on structured visual evidence point by point.
Signing blind vs. signing on per-point visual evidence of the die changeover
| Aspect | Classic die changeover | With iLEAN Edge + JIDOKA AI + SMED AI |
|---|---|---|
| Evidence the process lead signs on | Signed blind, with a risk of an out-of-tolerance start | Signed on per-point visual evidence, with the capture archived |
| Verification of bolster, die and alignment | By eye, per checklist, with no objective record | A CNN at each point compares against the reference state of the incoming part number |
| First-off validation | Manual and late — sometimes after the lot has started | Dimensional first-off confirmed before the start is released |
| Net SMED time | Extended by post-start readjustments after measuring the first stroke | −15 to −30% by avoiding post-start readjustments (estimate to be validated) |
| A camera that fails or cannot read | Assumed OK if nobody says otherwise | The block is held — absence of data is not a green light |
| 8D analysis when a changeover fails | Reconstruction from memory — nobody knows why it failed | Event documented in the evidence pack: photo + Edge verdict + first-off + signature |
Impact estimate for your plant — to be validated 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.
- Multi-reference Tier-2 press with several die changes a day, each one with its bolster, its die mounted per the order and its alignment as critical setup points.
- Edge + JIDOKA AI + SMED AI pilot on the die changeover (cameras plus start gate plus step-by-step guidance plus integration with the process lead's signature and the dimensional first-off). First value expected within 4 weeks.
- Estimated payback of 6 to 12 months, depending on the number of die changes per day and the weight of each part number in the mix. Estimate to be validated against your data.
- Return levers: direct PPM protection at the order change and improved press OEE through guided SMED, because changeover time drops between 15 and 30% once post-start readjustments are eliminated. Estimate to be validated.
- Protection of the relationship with the Tier-1, who penalizes out-of-tolerance starts under each OEM's requirements: every die changeover that starts in spec first time takes risk out of the supply relationship. Estimate to be validated.
And the fair question from the process lead
"What if the camera clears a die that is not actually the one on the incoming order, or the other way round, blocks the press for no real reason?" — hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI merely compares the current image of a critical point against its previously validated reference state — which is exactly what Edge does here — the best models brought the error below 1.5% [1]. And even then, the critical call is not made alone: JIDOKA AI holds the start, the first-off confirms and the process lead signs, seeing the evidence behind each point. iLEAN's three safety rings exist precisely for this — the AI proposes, the manager decides.
[1] OpenAI paper "Why Language Models Hallucinate", 2025 — on the reliability of AI in anchored tasks.
What people ask about the die changeover with Edge, JIDOKA AI and SMED AI
Why does the die changeover concentrate the PPM risk in a Tier-2 press?
Because every die changeover is a mini SMED that resets all the stamping conditions at once: bolster, mounted die, alignment and first stroke. On a multi-reference press that changes several times a day, that instant repeats often — and today it depends on a human signing "ready" without objective visual evidence point by point. If the start goes wrong, the first-off comes out of tolerance and the first lot goes into the Tier-1 container. PPM risk is not spread evenly across the shift: it concentrates at the order change, which is exactly the moment with the least structured evidence.
How does the camera verify the bolster, the mounted die and the alignment?
With fixed industrial cameras trained to recognize three things before the start: a clean bolster — no offcuts or scrap left from the previous reference —, the correct die mounted per the order and alignment OK. Each camera learns the reference state of that point for every part number in the mix and compares the current image against it at every change, not once per shift. It does not measure generically: it verifies that what is mounted is what the production order requires and that the bolster is in a condition to produce a first-off within tolerance.
What does JIDOKA AI do when it blocks the press start?
JIDOKA AI is the layer that turns the Edge cameras' verdict and the process lead's signature into a block on the press start. It takes its name from the classic jidoka principle — stopping at the anomaly rather than letting it advance — applied by AI to the die changeover: until it receives clearance from every camera (bolster, die, alignment) and the electronic signature on that evidence plus the dimensional first-off, it keeps the press held. The gate does not open because the schedule is tight or because "it has always been started this way"; it opens when the setup evidence is complete and signed. If a camera loses its reading, it fails closed: absence of data is not a green light.
How does SMED AI guide the changeover without penalizing OEE?
SMED AI guides the operator step by step through the die changeover, so that verification does not add to changeover time: it replaces the post-start readjustments that today inflate net SMED. In a classic changeover the press starts, the first stroke comes out, it gets measured, the misalignment is found, the press stops and gets readjusted — and that cycle can repeat. With Edge, every point is compared against its reference state while the team executes the external SMED operations, so the first start is already the good one. The starting estimate is a 15 to 30% reduction in changeover time from eliminating post-start readjustments — an estimate to be validated against your press data.
Does Edge replace the process lead's signature or back it up?
It backs it up, it does not replace it. Edge does not eliminate the human signature or the dimensional first-off: it reinforces them. iLEAN keeps three safety rings active in parallel — Edge visual verification of bolster, die and alignment; the dimensional first-off of the incoming part number; and the process lead's electronic signature on that evidence. What Edge contributes is that the signature stops resting on a checklist run through in a hurry between two references: it is given on structured visual evidence point by point, captured at the moment of the change and archived in the evidence pack. No ring depends on another to exist.
SMED AI applied to your stamping press — a 4-week pilot.
We work on your plant's real data, not ours. How long does an average die changeover take you? A pilot on one press, assessment with no commitment.
Request estimated ROI within 48h ‹ See all 12 metal stamping cases See automotive