In brief
Quality, in a small manufacturer, isn't decided on audit day. It's decided every day, in six moves nobody calls "quality": seeing a problem, knowing where it came from, handling it in time, heading it off, deciding on numbers, and closing the loop. A shop that does these six things well is solid, certified or not. A shop that neglects one is fragile exactly there, however thick its quality manual.
This guide is the map. Each section poses one of the six questions, says how a thirty-person shop can answer it without a full quality department, and points to the articles that go deeper. Read it end to end for the overview, or jump straight to the axis that worries you.
Why six questions, rather than a binder of procedures
I spent fourteen years in quality before I started building software, and I came to one conclusion: you can describe any quality system with six questions. Not six clauses of a standard — six questions an owner can ask themselves driving home.
Certification is binary: you're compliant or you're not. Useful, sometimes essential to sell, but it's a threshold, not a measure. Two equally certified shops can live in two different worlds — one endures its problems, the other sees them coming. What separates them fits on no certificate. It lives in how they answer, day to day, the six questions below. If a scale speaks to you, I've laid out elsewhere how to measure a quality system's maturity across these six axes — this guide is the practical version, question by question.
A word of method before we dive in: nothing that follows requires hiring. Almost all of it comes down to changing where information is captured, and when. That's the thread running through the guide.
Do we see problems, and fast? — Detection
The first question, and the most revealing. How many different people reported a deviation at your shop this year? If the answer is "three, always the same three", it isn't that your shop has few problems. It's that most never surface.
A system that sees nothing isn't a healthy system: it's a blind one. And the blindness almost always has the same cause — reporting a deviation is too cumbersome, or too frowned upon, for anyone to bother mid-production. So the lever isn't "raise awareness"; it's making reporting so easy it stops being an act of courage.
For more on what a system that detects nothing reveals — or hides — see the critical points you can spot at a glance in a quality system.
Do we know where each thing comes from? — Traceability
The second question. Is every non-conformity tied to the process that produced it, to the part, the lot, the measuring instrument? Without that thread, you accumulate incidents. With it, you accumulate knowledge.
It's the broadest axis, because it touches everything that produces evidence. The inspection report is its heart, and it's a more contested document than it looks: four people read it with four opposing interests. How it gets made decides its worth — and it often gets made in the worst conditions, which carries a cost nobody writes down anywhere: the hidden bill of machinist time.
Two technical bricks hold this traceability up. Drawing reading, first — why generic OCR fails on a dimensioned drawing and what a purpose-built tool does instead. Metrology, next: a report is only worth the instruments that produced its numbers, and managing that outside the spreadsheet isn't a luxury but the condition for the evidence to hold up at audit.
Do we handle things in time? — Responsiveness
The third question. Not "did you open a corrective action", but how much time passes between opening and closing — and above all, whether someone checks, later, that the action actually worked.
This is where the difference plays out between a shop that solves its problems and one that files them. An action closed with "operator made aware", with no effectiveness check three months later, isn't a corrective action: it's a folder made to disappear. The time lost reworking the same causes flows straight back into the hidden bill, and the discipline of closing out is built largely in a yearly exercise wrongly dreaded — the management review.
Do we act before it breaks? — Prevention
The fourth question, and almost always the weakest axis. How many of your processes have had a serious risk analysis? How many actions did you launch this year on a problem that hadn't happened yet?
Prevention is hard because it's invisible: nobody congratulates you for the breakdown that didn't occur. That's exactly where framed automation changes things — not by deciding for you, but by surfacing the weak signals a person alone won't catch in the flow. I've described what an ISO 9001 made accessible to a small shop through AI actually looks like, and above all what AI can do in quality — and what it must never do alone. The line between the two is the whole subject.
Do we decide on numbers? — Steering
The fifth question. Are your indicators current at the moment you make a decision, or reconstructed the night before the management review? A dashboard you fill in for the auditor isn't a dashboard.
Steering means deciding on facts rather than impressions — and that takes living numbers, not a cobbled-together annual report. It's the very point of the management review run without a consultant, where scattered data becomes acted-on decisions. And to know whether your steering is mature or improvised, go back to the maturity scale: steering is the axis that most sharply separates the shops that watch themselves work from those that merely exist.
Does the loop turn on its own? — Improvement
The sixth question. Do decisions taken in the management review become real actions, with an owner and a date? And does the same problem come back every six months under a slightly different label?
Continuous improvement isn't an intention, it's a mechanism — and that mechanism needs two things people underestimate. A person who pushes, first: the innovator you're looking for already works for you, and how you treat them decides everything. A system that doesn't retype itself, second: as long as the same information is re-entered in five tools, the loop leaks everywhere. That's the case for an ecosystem rather than silos, where the work order is entered once.
The foundations: where the data lives, how the tool is built
Two questions fit none of the six axes, yet they condition all six. Where your data lives, and whom you trust to hold it.
On the first: your customers' drawings are intellectual property under NDA, and where they're hosted is no distant legal footnote — what Quebec's privacy law changes concretely for a shop is worth understanding before you entrust anything to software. On the second: a trade tool isn't a pile of features, it's a pile of coded convictions. I've opened the hood on what it means to build trade software with framed AI, because an owner has the right to know what's under the hood before putting their shop on it.
Near you
The approach is the same everywhere, but the industrial fabric changes from one region to the next — and with it the demands of the buyers. I've looked at that region by region: Sorel-Tracy and its metallurgical hub, Drummondville and its do-everything manufacturing base, the Mauricie from hydroelectricity to aluminum, Greater Montreal and its aerospace machining, and Saguenay–Lac-Saint-Jean, machining for heavy industry far from the big centres.
In closing: where to start
Six questions, two foundations, one geography. If the map felt dense, that's normal — a quality system is dense. But you don't build it in one block. You build it by finding, among the six axes, the one where your shop is weakest, and tackling that first. It's almost always prevention or steering; it's rarely detection, which people wrongly believe they've mastered.
The good news, the one I repeat at every shop I visit, is that this work doesn't start with software or a consultant. It starts with a management decision: honestly looking at the six questions, naming the weakest axis, and giving it real working time rather than good intentions. The rest — the tools, the automation, the structure — comes after, and comes more easily once you know what you're trying to fix. If you read only one more thing, make it the principles all of this rests on.
The principles described in this guide are the ones that guided the development of Asterion Solutions, a suite of trade-specific tools built for small manufacturers who want to structure their quality without multiplying paperwork.