Superintelligence for Tailors: Alteration Tickets and Fitting Dates
A practical guide for tailor shops on structuring alteration tickets and fitting dates so pickup reminders and fit notes stay accurate, with a checklist.

Superintelligence for tailors comes down to one practical question: can an agent send the right pickup reminder and the right fit note for the right garment, every time, without anyone retyping the details? It can, but only when the alteration ticket and the fitting date are structured as records the agent can read, and when a person still owns any change to a promise made to a customer. This article explains how a tailor shop can structure both so reminders and fit notes stay accurate.
Key takeaways
- Treat each alteration ticket as one record with one garment, one customer and one promised pickup date.
- Write fit notes as measured observations and decisions tied to a ticket state, not as a general impression.
- Store every fitting as a dated event linked to the ticket, so a moved appointment never overwrites the history.
- Let the agent read tickets and fitting events and draft messages, but keep any change to a promised date with a person.
- Measure accuracy by comparing a sample of sent reminders with the record, not by counting messages sent.
Why reminders drift away from the garment
Most alteration errors do not start in the workroom. They start at the counter and on the phone. A customer asks to move a fitting, the seamstress adds a note in the margin of a paper ticket, and the new pickup date lives on a sticky note that never reaches the system. A person can absorb that mess because they remember the conversation. An agent cannot. It will send whatever the record says, quickly and to everyone on the list.
The failure patterns are predictable: a reminder for a garment that is still being pinned, a pickup date carried over from an earlier version of the work, and a fit note that describes a fitting that was cancelled. Each one is a data structure problem before it is a messaging problem, which is why the fix starts with the ticket and the calendar rather than with the wording of the message.
How should an alteration ticket be structured?
One ticket should describe one garment and one agreed piece of work. A jacket that needs a sleeve shortening and a waist take-in is one ticket with two requested changes, not two tickets that happen to share a customer name. Each ticket should hold these fields:
- A ticket identifier that appears on the physical tag, the workroom bin and the message log.
- The customer contact and preferred channel, recorded once and linked to the ticket.
- The garment as received: type, colour, label details, and any existing damage noted at intake.
- Requested changes, one per line, written as actions (take in, shorten, replace zipper) rather than as wishes.
- The current state, chosen from a short fixed list: received, pinned, in work, ready, collected.
- The promised pickup date, plus a separate field for the date the shop last confirmed it with the customer.
- An open-question or decline note for anything the shop has not yet agreed to do.
The fixed state list matters most. A reminder rule can say that a pickup message goes out only when the state is ready and the promised date is confirmed. Free-text status cannot be checked by a rule, so it cannot be checked by an agent either. Shops that run repairs on a similar ticket pattern can compare the approach in AI for shoe repair shops.
Treating fitting dates as events, not fields
A fitting is something that happens on a specific day, with a specific person, at a specific stage of the work. Storing it as one date field on the ticket hides that. When a second fitting is moved, the old date gets overwritten, and nobody can tell whether the fit notes belong to the first appointment or the second. Keep fittings as separate events linked to the ticket. Each event carries its date and time, the stage it belongs to, the shop staff member present, the outcome, and whether another fitting is needed.
When a fitting moves, close the original event with a reason, such as customer request or workroom delay, and open a new one. The history stays visible, which helps when a customer asks why a hem was checked twice. Ski rental shops face a similar problem with fitting records, and the same separation of events from fields applies. Agents should read the open event when they decide what to tell a customer, never the closed one.
Writing fit notes an agent can rely on
Fit notes are where accuracy usually breaks down, because they are written in a hurry and read later by someone who was not in the room. Phrases such as looks good or a bit tight are hard to act on. Write what was observed and what was decided, in short sentences tied to a specific change on the ticket:
- Observation: the left side seam sits wider than the marked line at the waist.
- Decision: take in the left side seam to the chalk line, then re-pin before sewing.
- Next step: second fitting at the pinned stage, with the customer present.
- Author and time: staff initials and a timestamp on every note.
Keep observation and instruction in separate fields. An agent may summarize an observation for a customer, but it should never turn an open instruction into a promise about the finished garment.
Illustrative scenario: a customer books a second fitting for a wedding guest dress. The day before, the seamstress calls to move it from Thursday to Friday because the lining arrived late. If the Thursday event is still open in the calendar, the reminder tells the customer to come in on Thursday, and the customer arrives to find the dress still on the bench. In the structured version, the Thursday event is closed with its reason, the Friday event is open, and the agent reads only the open event before drafting anything. The person who made the change still decides whether the new date goes out.
Pickup reminders that follow the ticket state
A reminder rule should be short enough that a front desk employee can read it aloud. A workable plan looks like this:
1. Export the open tickets and fitting events into one list that shares the ticket identifier. 2. Check that every open ticket has one of the five states and that every ready ticket has a confirmed pickup date. 3. Write the reminder rules in plain language, for example: send a pickup reminder only when the state is ready and the confirmed date falls within the coming week. 4. Run the agent in draft mode for a period, and have a staff member compare each draft with what the shop would have sent. 5. Turn on sending only for the reminder types that match the record in the review sample, and keep drafts for any message involving a changed date. 6. Give every outgoing message a unique key tied to the ticket and the reminder type, so a retried job cannot send the same reminder twice.
The last step borrows a standard pattern from payment systems. An idempotency key tells the receiving service that two requests are the same operation, so a retry does not become a duplicate. Stripe documents how this works for its API, and the same idea applies to a customer message that a queue might resend after a timeout.
Where the agent stops and a person decides
A short decision guide keeps the boundaries visible to everyone in the shop:
- Draft and send: a standard pickup reminder for a ticket in the ready state, with a confirmed date and no open question.
- Draft for staff review: any change to a pickup date, any request to move a fitting, and any reminder for a ticket with a decline note.
- Hand off without drafting: complaints about fit or damage, requests for refunds or price changes, and anything about a garment the shop has not yet received.
- Never: promise a finished date that is not in the record, or tell a customer a garment is ready while a fitting is still open.
Permissions should follow the same table. The agent gets read access to tickets and fitting events, write access only to drafts and its own log, and no ability to edit the promised date. The OWASP authorization guidance states the general principle: grant each identity only the access its job requires, and check that access on every action rather than once at setup.
How Be Human approaches this: in a first engagement, Be Human defines one role for the shop, such as a pickup and fitting coordinator, and writes down the memory that role may use, which here means the ticket record, the fitting events and the reminder rules. The agent receives the tools it needs to read the calendar and draft messages, the approval boundaries described above, and one observable success criterion, such as every reminder in a weekly sample matching the ticket state. Managed operation then covers monitoring, permission review, incident response when a reminder reaches the wrong person, and controlled changes when the shop adds a service. The shop keeps control of its records and of its customers.
Keeping the record accurate after launch
Accuracy is a habit, not a setup step. Each reminder the agent sends should be logged with the ticket identifier, the state it read, the time it read that state, and the rule that fired. This is the kind of trace the OpenTelemetry observability primer describes for software systems in general: enough detail to follow one action from start to finish. When a customer says a reminder was wrong, staff can see exactly which record produced it.
Review on a fixed rhythm. Sample a handful of reminders each week and compare them with the tickets and fitting history. Log every mismatch as an incident, fix the record or the rule, and note the cause, whether it was a phone change that was never written down, a state that was not updated, or a rule that read the wrong field. Review agent permissions whenever the shop adds a service or a staff member, and remove access that no longer matches the job. When the list of states grows, update the reminder rules in the same change, not later.
If you are not sure which part of your shop should go first, take the free Be Human AI diagnostic to find the first task worth handing to an agent.
Frequently asked questions
Does superintelligence for tailors mean replacing the people in the workroom? No. In a well-structured shop, the agent handles the messages that follow from records, such as reminders, confirmations and status questions, so people spend less time chasing details. Fitting, cutting and the judgment about how a garment should hang stay with the tailor. The agent's role is to keep the record that the tailor and the front desk share accurate.
Which tickets should an agent handle first? Start with tickets already in the ready state that have a confirmed pickup date. These carry the least ambiguity, and a mistake is easy to check. Leave tickets with open questions, customer complaints and garments with prior damage with staff until the reminder rules have run cleanly for a while.
Can an agent change a fitting date by itself? It should not. An agent can read the open fitting event, draft a message proposing a new time, and log the request. The change itself belongs to the person who owns the calendar, who closes the old event and opens the new one. Writing that boundary down makes the record easier to trust for everyone who uses it.
How do we know the reminders are accurate? Compare a sample of sent reminders with the ticket state and fitting history on a fixed schedule. Keep a log that shows which record and rule produced each message, so any mismatch can be traced. If the sample reveals a pattern, such as reminders sent after a phone change that nobody recorded, fix the intake step first rather than the wording of the message.
Your next step
See where an AI agent fits in your company first.
Take the free Be Human AI diagnostic. A few questions about your work show which task to hand to an agent first, what it needs to know and where a person approves.
Be Human editorial guidance. Examples are illustrative unless explicitly attributed. Sources provide background; implementation choices depend on your workflow.


