AI Measurement in a Tailor Shop: What the Fitter Still Does

The shop in Lyon had been open for eleven years and the owner still wrote every order on a carbon-copy pad. Cloth reference in pencil, measurements in pen, customer name at the top, date at the bottom. He posted it to his factory by mail. When he asked me what would change if he moved to an online ordering system, I told him the truth: the pad goes away, but the tape does not. Tailoring software handles the part of his job that a notebook can handle. It does not handle the part that requires him to stand in front of a customer and read what no form can ask.

That distinction matters more now than it did five years ago, because the question tailor shops ask us has shifted. It used to be "which factory should I work with." Now it is "which software do tailor shops use to place MTM orders, and do I still need to measure." The answer to the first part is straightforward. The answer to the second part is the subject of this piece.

What tailoring software actually handles

The ordering pipeline is the part that belongs to the screen. On our platform, a shop logs in, starts an order, selects the customer, picks cloth and buttons, enters measurements, chooses style options, reviews the whole thing, and submits. Eight steps in a browser, in English, Chinese, French, Spanish or German. After that the shop tracks the order and messages the factory in the same place.

That pipeline used to live on paper, in spreadsheets, or in email threads that nobody could search six months later. What software does well is hold all of it in one spot: the cloth reference, the measurement set, the style choices, the order status, the conversation with the factory. It eliminates double entry, because the numbers the fitter types are the numbers the cutter reads. It eliminates the lost fax, the misheard phone call, the measurement sheet that went through the wash in somebody's trouser pocket.

What it does not do is decide what those numbers should be.

What the AI suggests and where it stops

CENTI AI takes two inputs, the customer's height and weight, and returns a suggested size for every line on the order form in about 30 seconds. The suggestions appear in grey. Type a measured value over any of them and the measured value is used. Leave a suggestion in place and the workshop cuts to it as written.

That is the full scope of what the AI does. It does not scan anyone. It does not photograph anyone. It works from two typed numbers and a statistical model, and it gives the fitter a starting point rather than a finished answer.

The starting point is useful. For a customer who is not in the shop, or for a first order where the shop wants to see how close the algorithm lands before investing fitting time, height and weight are enough to place an order. But for any customer the fitter can reach with a tape, the AI suggestion is a draft, not a decision.

Three ways to get a size exist on the same order, and they can be mixed:

1. Height and weight alone. Enough for a first remote order.
2. A garment that fits. The customer measures a jacket or trousers they already like.
3. A tape on the body. In the shop, or at home with video guidance.

The platform accepts all three and treats the most specific input as the one that counts.

The five things the fitter still decides

Here is the short list of what no ordering system handles, and what a shop's fitter is actually being paid to judge.

1. Shoulder width. The single most critical measurement on a jacket. Our system runs to 47 measurement fields, 31 of them required, and shoulder width is the one field with no tolerance worth having. Move it a centimetre past the bone and the sleeve head collapses. An algorithm working from height and weight gives you a population average. The fitter gives you the bone.

2. Posture. A customer who stands with his shoulders rolled forward needs a different back balance than one who stands upright. The order form has a field for it. The AI does not fill it. The fitter reads it from across the fitting room, not from a number.

3. Asymmetry. Most people have one shoulder slightly lower than the other, or one hip slightly wider. An algorithm trained on averages sees neither. A fitter standing behind the customer with a tape sees both.

4. Drape preference. "I want it slim" means different things to different people. The fitter translates the customer's language into a wearing style (slim, standard, or relaxed), a waist shape (flat, standard, or round), and a chest size. Those three choices set the block before the workshop touches the cloth, and they are judgement calls, not calculations.

5. What went wrong last time. The customer who says "my last jacket pulled across the back" is giving the fitter a correction that does not exist in any database. The fitter decides which measurement to adjust and by how much. The software records the correction. It does not make it.

Remote orders: the tape is still someone's job

The question that follows is always the same. If the customer is not in the shop, who does the fitting?

The customer does, with guidance. On our measurement page at b2b.centisartoria.com/measure, the customer enters name, height, and weight and gets a starting set. A short video shows where the tape goes for each body part. The diagram shows where to measure. It is not a scan. At the end the customer receives a measurement code, sends it to the shop, and the shop imports it into the order without retyping anything.

The tape is still someone's job. The difference is that "someone" can now be the customer in another city, following a video on a phone screen, rather than the fitter in the shop. The fitter reviews the imported numbers before submitting and decides whether to trust them or ask for a reshoot. For a first order with a remote customer, many shops start from height and weight alone, check the garment when it arrives, and refine on the second order.

The minimum for any of this is one suit. A two-piece starts from USD 100 with the shop's own cloth (CMT) or USD 155 with ours, sourced from 33 mills and cloth merchants, 26 of them in Europe. Production takes 14 days from a confirmed order, shipping on top.

The software and the fitter are not competing

The carbon-copy pad in Lyon held the same information our platform holds. What it could not do was check itself, carry forward a returning customer's measurements from the last order, or show the shop whether the factory had started cutting. The software replaced the pad. It did not replace the man standing next to the mirror with a tape around his neck and one eye on how the customer shifted his weight when he thought nobody was looking.

If your shop is choosing tailoring software for the first time, the question is not whether it replaces the fitter. It is whether it frees the fitter to do the five things listed above instead of recopying numbers into a spreadsheet and chasing an order status by email. The answer, from every shop we have seen make the switch, is yes. The fitter measures better when the admin is off his desk.

We stand behind every piece: if something is not right, we make it right. See prices and your margin, 3 min, or send us your shop name and we will set up a platform login.