Industrial AI in ophthalmic lens and frame factories
Eyewear manufacturing has a peculiarity no other sector has: the piece going out the door is at once a fashion accessory and a medical device. It has signed optical tolerances, a medical device quality system, a mandatory impact test with records that must be kept for years, and on top of that an end user who looks at the lens against the light and returns the pair if they see a speck.
And it has a second peculiarity that changes everything: there is no reprocessing. A prescription lens badly surfaced, badly coated or badly mounted is not recovered. It is remade. That is why the indicator summarizing an optical plant's health is not productivity, it is the remake rate — and the remake rises from causes living in different places: a prescription change that arrived late by email, typing at the edger, a vacuum load that drifted with nobody watching, a different cosmetic criterion on the night shift.
This page holds the twelve industrial AI use cases we have landed on this sub-sector, grouped by block of the IRIS architecture: capturing what is lost today (Connect), seeing the line in real time (Edge) and turning memory into decisions and dossiers (Agents). None of them replaces anyone: all start from the person staying in command and signing.
Capturing what is lost today
The surfacing shift report, paperless
Payback 4-9 months · from 8-24 h of latency to second zero
See the case →
The old vacuum chamber starts to talk
Payback 5-10 months · replacement CAPEX deferred
See the case →
The prescription change arriving by email
Payback 3-7 months · from 2-4 h of latency to seconds
See the case →
The GEMBA walk that saves itself
Enabling piece: the GEMBA stops dying at the end of the shift
See the case →
Two-tap human verification
Enabling piece: nothing enters the master without a human signature
See the case →
From the laboratory's Excel to the ERP
Payback 4-9 months + −50/−80% quality module licenses
See the case →
Receiving without typing the delivery note
Payback 3-8 months · truck from 45 min to a few · −50/−70% licenses
See the case →
The prescription reaches mounting on its own
Payback 3-6 months · zero typing in special jobs
See the case →AI that sees the line
Tell us where it hurts in your optical plant.
We work on your plant's real data, not ours. Assessment with no commitment.