In brief
Sorel-Tracy did not grow up making small production runs. It grew up pouring steel, building heavy fabrications and assembling hydroelectric equipment — turbines, generators, gates — for Quebec's great dams. That heritage shaped a particular industrial culture: the culture of the one-off, massive, expensive part, where a mistake cannot be undone in five minutes.
I spent part of my fourteen years in quality inspecting exactly that kind of heavy equipment, including hydroelectric work. It is that perspective — the discipline a part you cannot remake demands — that I want to bring to bear on the region's manufacturing fabric, and on what quality has really become here.
A region built on the big part
You don't need to introduce Sorel-Tracy to people in the trade. It is one of Quebec's cradles of steel and metallurgy: steel, castings, heavy fabrication, the machining of parts few shops elsewhere would dare put on a table. But reducing the region to its steelmaking misses half its story.
Sorel-Tracy was also a hub for the manufacture of heavy hydroelectric equipment. When Quebec electrified its territory, someone had to machine and assemble the turbines, generators, gates and massive components that equip the great dams. A share of that work happened here, in the region's shops and industrial yards. It is an industrial heritage few people spontaneously associate with the town, yet it explains a great deal about its technical culture.
Around those large customers orbits a whole network of small and mid-sized manufacturers — mechanical subcontracting, machine shops, fitting, industrial repair — who live off their standards. And those standards have a very specific nature: they bear on parts that matter, not on volumes that forgive.
A turbine, a big heavy part — it doesn't get remade because someone wrote a dimension down wrong.
What hydroelectric work taught me about quality
Early in my career I had the chance to do quality control on hydroelectric equipment. It is a school of its own. When you inspect a part meant to turn for decades at the heart of a dam, you are no longer in the logic of sampling where a few defects are tolerated. You are in the logic of proof: every critical dimension has to be taken, recorded and tied to a record a third party can verify years later.
What I took away from those years is that on the big part, the inspection report is not paperwork. It is the deliverable. A part without its conformity file is not really delivered. ISO 9001 clause 8.6 on release of products says exactly that: nothing leaves until there is proof the requirements are met. In hydroelectric work, that rule is not an administrative formality — it is a safety condition.
That mindset shows up in Sorel-Tracy's machine shops today. When you machine a heavy part, a one-off or a very short run, for a major customer, rework is expensive, the schedule is critical, and the customer wants documented proof that what they receive conforms to the drawing they handed you. A well-regarded shop in the region put it to me simply: "the customer doesn't just pay me for the part, they pay me for the report that comes with it."
Inspecting heavy means proving heavy
Machining big also means inspecting big. A large part often means many dimensions to take, tight tolerances at critical spots, varied instruments, sometimes several operators on one file. First-article inspection — validating the first part of an order before releasing the rest — becomes a pivotal moment: it either authorizes the rest of the run or holds it.
I have seen how that inspection is still too often done in the region's shops: the dimensioned drawing on the table, the caliper or micrometer on one side, and at the end, someone hand-copying every dimension into a spreadsheet or a form. The inspection itself is done seriously. It is the transcription that costs: recopying dozens of values, formatting them, generating the report, and starting over on the next file.
In a shop machining high-value parts, that time is qualified inspector time — one of the most valuable and scarcest roles on the floor. How many hours a month do your best inspectors spend recopying numbers they just measured, instead of inspecting the next part? The question deserves an honest answer, because in many shops it is higher than people think.
Metrology, the blind spot of the big part
Here is another thing hydroelectric work drove into my head: an inspection report is only worth something if the instrument that took the measurement was itself valid at the time of the check. Easy to say, harder to prove eight months later, when a customer disputes a part.
ISO 9001 clause 7.1.5 on monitoring and measuring resources asks for nothing less: knowing which instruments are used, that they are calibrated, and being able to demonstrate it. Too many shops still manage this in a spreadsheet that ages badly, where a calibration due date slips by unnoticed. On a critical part, that is exactly the kind of gap that turns a dispute into a nightmare: the part may well have been good, but you can no longer prove it.
What rigorous quality changes
A good tool does not replace the inspector. It takes away the thankless part. A well-built trade-specific software reads the dimensioned drawing, extracts the dimensions and tolerances, and prepares the report instead of the person who would otherwise retype them one by one. The inspector takes the measurements, validates, adjusts what needs adjusting — and signs. The deliverable comes out structured, consistent, ready for the customer.
Yes, there is AI behind that kind of extraction. But let's be honest about its role: AI on its own is not reliable — it proposes, it guarantees nothing. What makes it useful and safe is being framed by a structured software that imposes the shape of the report, keeps track of who validated what, and refuses to let an incomplete file through. Framed properly, AI does the bulk of the repetitive work; the judgment — is this out-of-tolerance dimension acceptable, should a nonconformity be opened — stays entirely human. The inspector signs, never the machine.
That frame has a decisive effect on the big part: nothing gets lost in silence. When a dimension is out of tolerance, it does not vanish into a mis-saved spreadsheet. It becomes a nonconformity (clause 8.7), attached to the process that produced it. If it calls for a fix, that fix becomes a corrective action (clause 10.2) attached to the nonconformity. The chain is traced, from drawing to decision — and an auditor, like a demanding customer, can follow it.
In closing: what if the report worked for you?
Sorel-Tracy carries in its shops the heritage of the turbine and the heavy part. That mechanical know-how has nothing to prove. The real question is elsewhere: in the time quality costs in transcription, and in the ability to prove, file after file, that the part delivered truly conforms to the drawing received — the same demand, at bottom, that I knew on hydroelectric equipment.
So ask yourself honestly. How many qualified-inspector hours go every month into recopying dimensions? Could you, tomorrow, retrieve the calibration record of the instrument that validated a part delivered eight months ago? And what if the report, instead of being the end-of-file chore, became the automatic by-product of an inspection done right?
The principles described in this article are the ones that guided the development of Asterion Solutions, a suite of trade-specific software designed for manufacturing SMEs that want to structure their quality without multiplying administrative tasks.