In brief
Greater Montreal is one of the few places in the world where a complete aircraft can be assembled without leaving the region. That aerospace cluster pulls up a whole network of precision-machining shops — engines, landing gear, structures, critical parts — and with it, a level of documentary rigour few markets impose.
Here, the first-article inspection is no formality: it's an AS9102, sometimes a full PPAP package, with tight tolerances and a traceability trail that must survive a customer audit. After fourteen years in quality, I've settled on a simple conviction: in this region the real question isn't "how do we machine it" — the shops know — but "how do we prove it, fast and flawlessly."
A region that learned to manufacture rigour
It helps to remember how this cluster was built. Montreal didn't become a world aerospace hub by accident. Since the postwar years, the region has accumulated engine makers, aircraft builders, systems and equipment suppliers, landing-gear specialists. Around those primes, decade after decade, a dense network of precision machine shops took shape — from the South Shore to Laval, from the West Island to the northern ring.
What strikes me, looking at that fabric, is that it organized itself around a culture before it organized itself around a product. The culture of proof. Many of these shops don't build the aircraft; they build the parts the aircraft cannot afford to get wrong. That responsibility forced them, very early, to document their work as few other sectors demand.
And aerospace isn't alone in this ecosystem. The region also machines for medical, rail, defence, energy. Different markets, same instinct: the part isn't enough, you need the file that certifies it. A Montreal shop that wants to climb the value chain always gets caught by the same reality — the customer wants formal proof, structured and auditable.
In this region, you don't just sell a conforming part. You sell the demonstration that it is one.
AS9102, PPAP: the real weight of documentary conformity
Take the aerospace first-article inspection, the AS9102. In principle, nothing exotic: before launching a production run, you prove that the process yields a first part perfectly conforming to the drawing. In practice, it's a punishing exercise. Every characteristic on the drawing must be identified, numbered, tied to its measurement method, its result, the instrument used. Nothing can be missing. One forgotten characteristic, and the form is rejected.
PPAP, inherited from the automotive world and increasingly requested in the region, goes further still: it stacks control plans, capability studies, material records, samples. It isn't a document, it's a file — sometimes a full binder for a single part number.
Yet in many shops this work is still built by hand. You reopen the dimensioned drawing, renumber the characteristics one by one in a spreadsheet, recopy the tolerances, calculate the minimums and maximums, prepare the measurement columns. On a dense aerospace part, that gymnastics repeats dozens, sometimes hundreds of times — before an instrument has ever touched a part. It's expert time consumed by transcription. And it's precisely there, in the recopying, that the errors slip in — the ones that get the file rejected.
The standard leaves no room for improvisation. Release of products (clause 8.6) requires that the planned verifications be carried out and documented, with a record of who authorized the release. No conforming file, no shipment. In this trade, the document is literally part of the part.
What I learned to look at first
When I assess a shop, I never start with the machine. I start with the path the proof travels. Where a dimension leaves the drawing, whose hands it passes through, how many times it's recopied before it lands in the file delivered to the customer. That's where the fragility hides — in those transfers, not in the cutting of the metal.
Because a Montreal shop can have flawless machining skill and still lose a contract over a badly built AS9102. The part was good; the file didn't prove it. I've seen that scene often enough to turn it into a rule: the quality of the demonstration counts as much as the quality of the execution. And that demonstration rides largely on repetitive tasks people dislike — and therefore do poorly, not through incompetence but through fatigue.
It's that repetitive layer — reading the dimensions, building the table, numbering the characteristics, calculating the tolerances — that a well-designed trade-specific software can absorb. Automatic extraction of dimensioned drawings, framed by strict rules, takes on the tedious part. The inspector keeps the judgment: choosing the instrument, measuring, ruling on an ambiguous dimension, signing the release. The tool prepares the bulk of the file; the judgment stays entirely human.
The nuance is critical, and I repeat it because it's the heart of everything. Extraction left to itself fails in silence: it returns a plausible value with the same confidence whether it's right or wrong. On an AS9102, a misread but credible dimension is far worse than an obvious blank. What makes the tool reliable isn't the magic of automatic reading — it's the software frame around it: making doubt visible, flagging the dimensions to review, tying each balloon to its line by a deterministic rule rather than a guess. Extraction without that frame is a stage set that stands until the audit, then collapses.
Metrology: the blind spot that costs dearly
In an aerospace file, every measurement is only worth as much as the instrument that produced it. If the micrometer or the measuring column wasn't calibrated — or its calibration lapsed — every value in the file loses its evidentiary weight. The standard is clear: monitoring and measuring resources (clause 7.1.5) must be calibrated, verified, their status known at all times.
In many shops, this still lives in a spreadsheet someone updates "when they think of it." It holds until the day a customer auditor asks for the calibration certificate of the instrument used for a given dimension, on a given serial number. In aerospace, that question isn't hypothetical: it comes. A coherent software ecosystem links the instrument to its calibration record, warns before the due date, and stops a lapsed instrument from slipping into a report. The inspection report and metrology aren't two files — they're two ends of the same chain of proof.
Proximity, and control of your own drawings
The drawings of an aerospace part, the history of your inspections, your first-article files: these are sensitive assets, often covered by strict non-disclosure agreements with your customers. Knowing where that data lives, who controls it and who owns it isn't a detail — it's a requirement your customers will pass down to you sooner or later. Processing and hosting in Canada gives you a simple, verifiable answer: sovereignty here means control and ownership, and your drawings stay drawings you command. It's worth being precise about what the law protects, too. Quebec's Loi 25 — the province's private-sector privacy law — and its federal counterpart PIPEDA protect personal information, such as the names of your inspectors and machinists. The drawings and designs themselves are intellectual property, protected by NDAs and trade-secret practice, not by privacy law. Both matter; they aren't the same thing.
And then there's plain proximity. I come from the shop floor, I speak the language of the plant, and I answer in French as readily as in English — because a Montreal supplier works with customers in both languages every day. Understanding the local ground isn't a marketing line: it's what makes a tool actually fit the way people work in Greater Montreal.
In closing: would your file survive the audit?
Greater Montreal's precision machine shops don't lack know-how — they overflow with it. What holds them back is the invisible weight of proof: the expert hours swallowed by transcribing an AS9102, the forgotten characteristics that get a file rejected, the calibration certificates hunted down at the last minute before an audit.
So the questions I leave you with answer themselves, with your own numbers. On the build time of your next first article, how much is judgment, and how much is mere recopying? How many of your recent nonconformities came from a part, and how many from a mis-recorded value? And if a customer demanded tomorrow the complete file of a part you shipped three months ago, how long would it take you to produce it — with no gaps?
The principles described in this article are the ones that guided the development of Asterion Solutions, a suite of trade-specific software built for manufacturing SMEs that want to structure their quality without multiplying administrative work.