For working tailors
AI tools for tailors: what actually helps a shop
Almost everything written about AI and clothing is aimed at brands with a technology budget. For a shop of one to five people the useful list is short, and most of it has nothing to do with fitting. Four jobs AI does well today: pulling rough measurements from photos, keeping up with client follow-up, producing decent visuals and social posts, and clearing admin. Three it does not: judge a body, run a fitting, or decide cloth. This page is the honest split. Written by a production partner that receives the orders these tools generate, and sees which ones come through clean.
The tools worth a shop's time
| What it does | Where it earns its keep | Where it falls down | |
|---|---|---|---|
| Photo & scan measurement | Extracts body measurements, or recommends a size, from photos or a phone scan | Remote and travelling clients; a starting set of numbers before the client is in front of you | Posture, asymmetry and stance are exactly what it misses. The things that decide whether a jacket sits |
| Client follow-up | Drafts messages, reminders and quote replies from a short prompt | The highest-value use for most shops: orders are lost to silence, not to price | Reads as a template unless you rewrite it in your own voice; never let it invent a delivery date |
| Visuals & mock-ups | Renders cloth, colour and style combinations before anything is cut | Settling a client's indecision at the counter, and making a swatch look like a finished jacket | A render is not a sample, showing one as if it were a made garment sets up a complaint |
| Social content | Cleans up product photos, removes backgrounds, drafts captions | Shops that win clients on Instagram; it removes the excuse of having nothing to post | Generic captions read as generic; the craft detail that earns the follow still has to come from you |
| Admin & books | Reads receipts and invoices, sorts orders, drafts documents | Getting an evening back each week in a shop with no office staff | Needs checking against the real numbers; confident wrong totals are the classic failure |
| Translation | Live translation of client and supplier messages | International clients, and sourcing cloth or production abroad | Trade vocabulary goes wrong, construction and fitting terms are worth confirming |
Named vendors change quickly in this category, so evaluate against the questions above rather than a brand. Widely used measurement tools include 3DLOOK, Bodygram and MySize; general assistants such as ChatGPT and Claude cover the drafting, admin and translation rows.
What the vendors actually mean
Five distinctions worth knowing before you pay for anything.
- Measurement extraction vs size recommendation
- The single most useful distinction, and the one most vendor pages blur. Extraction returns numbers, chest, waist, sleeve, that a pattern can be cut from. Recommendation returns a size label such as 40R, which is only useful if you are selling ready-made or a fixed size grid. A shop cutting to measure needs the first; much of what is marketed as AI fitting is the second.
- Photo-based measurement (2D)
- The client stands against a wall and takes two photos, usually front and side, with a height reference. Software estimates girths and lengths from the silhouette. Good enough as a starting point for a remote client; sensitive to posture, footwear, clothing bulk and camera height, which is why the same person can measure differently twice.
- 3D body scan
- A phone sweep or booth captures a body mesh rather than two flat images, so girths are measured rather than inferred. More consistent than 2D, and still blind to how someone habitually stands. A dropped shoulder reads as a dropped shoulder, but whether to correct it or cut around it is a tailor's judgement.
- Hallucination
- When a general AI assistant produces something fluent and wrong. An invented lead time, a fabric composition it never had, a total that does not add up. It matters commercially: anything a client could hold you to, especially dates, prices and compositions, has to be checked before it is sent.
- Digital measurement handoff
- Passing a measurement set from whatever tool produced it straight into a production order, rather than retyping it. The practical test of any tool is whether its output can be handed to a maker in a form they accept without a phone call. Centi Sartoria takes measurements from any source, including a set the client's own app produced.
Questions shops actually ask
Can AI take a client's measurements from a photo?
It can produce a usable starting set, chest, waist, hip, rough lengths, from two photos with a height reference. Treat it as a first draft rather than a measurement sheet. What it cannot see is posture, shoulder slope, stance and asymmetry, and those are what decide whether a finished jacket sits correctly. For a client you will never meet, it is far better than guessing; for one standing in your shop, your tape is still faster and better.
What is the difference between a size recommender and a measurement app?
A recommender outputs a size label like 40R by matching a body to a size chart. A measurement app outputs actual numbers a pattern can be adjusted from. Recommenders are built for retailers selling fixed sizes; if you cut to measure, they give you nothing you can use. Check which one a vendor is selling before paying. The marketing language for both is nearly identical.
Can I place a production order using AI-generated measurements?
Yes, and shops do. Centi Sartoria accepts measurements from any source. Your own tape, a client's scanning app, or a size set. Because the numbers are what production needs, not the method that produced them. The usual approach for a remote client is to order one garment first, note what the fitting reveals, and carry those corrections into every later order for that client.
Will AI replace tailors?
It has not replaced the parts that are hard. Reading a body, deciding what to correct and what to leave, choosing cloth for how someone actually lives, and the fitting itself are all judgement work, and nothing on the market does them. What AI has genuinely taken over is the surrounding load, writing, chasing, photographing, filing, which is where most small shops lose their evenings.
Which tool should a one-person shop pay for first?
A general assistant for messages, quotes and captions, before anything fitting-related. Follow-up is where orders are lost, and it is the cheapest thing to fix. Measurement tools are worth adding when remote clients become a real share of the business, not before. Until then you are paying to replace something you already do better.
Is it worth digitising measurements if the shop is small?
The value is not the software, it is that a client's corrected measurements survive. Kept properly, a repeat order becomes a message rather than an appointment, which is the single biggest change to a small shop's economics. A spreadsheet does this. The tool matters much less than the habit.
The part AI does not solve
Every tool above ends at the same place: someone still has to cut and make the garment. Centi Sartoria is a Shanghai atelier of 200 house tailors that produces made-to-measure suits, jackets and coats under your own label, from a one-piece minimum, in about 14 to 20 working days. Measurements accepted from any source.