Wet film thickness control in coating — measured while you apply, not discovered at dry film.

The problem

Dry film is measured when nothing can be fixed — and wet film is measured but never recorded.

Wet film thickness × volume solids = predicted dry film thickness. In other words: the wet reading is the only one that arrives while the coating can still be corrected. And yet, on most lines, the control story looks like this:

1. The comb gauge is used, but the reading is never recorded — the applicator checks, adjusts and moves on; the number lives and dies in his head.
2. Dry film is measured hours later — after flash-off and the drying or oven cycle. If it comes in low, repainting means another full cycle; the mistake is already cured onto the part.
3. Overthickness triggers no alarm at all — “plenty of coverage” passes every check, while product is given away run after run.
4. Viscosity and thinning drift between batches and shifts — and the compensation lives in the veteran applicator's judgment, written down nowhere.

The line lead knows it; the quality manager knows it. The classic control works most of the time. What it can never do is arrive early — by the time dry film is measured, the choice is repaint or absorb the cost.

How it fits the IRIS system

iLEAN doesn't replace your gauge or your paint — it captures the reading you already take and ties it to the order.

Thickness control does not fail for lack of a gauge; it fails because the wet reading is never recorded or tied to the order. iLEAN acts as the putty between the comb gauge, the product datasheet and the production order, changing none of the three.

Edge identifies the order and the zone. Connect captures the wet reading. The agent computes the window from volume solids and warns before the film dries. The person signs — never the other way round.

The iLEAN pieces applied to wet film control:

• Edge — a terminal on the line. It identifies the order and the product by QR and, by voice or camera, the zone of the part being measured. An applicator with an earpiece says “wet ninety” and the reading lands tied. It works without a network.
• Connect — if the applicator prefers a photo, it reads the last wetted tooth of the gauge from the image; if the line has non-contact measurement (ultrasonic, laser), it captures the value straight from the instrument. It also takes volume solids and declared thinning from the product datasheet.
• Window agent — computes the wet film window from the specified dry film thickness and the volume solids, checks every reading and warns the applicator at second zero if it drifts out — while the film is still wet. An order with out-of-window readings, or none at all, is held until quality signs. The person signs — the order never closes on its own.

See the full IRIS architecture →

Before and after

Unrecorded comb gauge vs. wet film tied to the order with iLEAN

Aspect	Comb gauge + dry film checks	With iLEAN Edge + Connect (wet)
Wet reading	Taken and forgotten	Captured, tied to order and zone
Low film build	Discovered at dry film, hours later	Warned during application, still wet
Overthickness	Invisible: “plenty of coverage”	Quantified per order and part number
Thinning and viscosity adjustment	Veteran judgment, unwritten	Pattern captured per part number
Dossier per order	Final dry film thickness only	Wet + dry + application conditions
Repaints	Another full cycle per low part	Falling: corrected while wet

Impact estimate

Impact estimate for your plant — to be validated with your numbers.

The block below is an estimate to be validated with the specific data of your plant. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.

• Liquid coating line (industrial metal, plastics, wood, coil) with a dry film specification per product and a comb or wheel gauge available.
• Edge pilot on one line (terminal + voice or photo capture + wet-window agent). First value expected within a few weeks.
• Indicative payback between 5 and 10 months, depending on the cost of the applied product, line volume and the current repaint rate.
• Hard levers: ≥ 25% reduction in repaints for low film build (estimate to be validated); overconsumption from overthickness quantified and falling; a dossier per order with wet + dry + conditions.

And the quality manager's reasonable doubt

“What if the system misreads a value?” — false positives exist, and that is why the system never decides alone. Hallucination is a problem of free generation, not of anchored tasks. In tasks where the AI merely receives a thickness reading and compares it against the window computed from the datasheet volume solids, the best models brought error below 1.5% ^[1]. And even so, what is critical is never decided alone: iLEAN warns and the operator or the quality manager sign. The three safety rings exist precisely for this.

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