AI for Dance Studios: Class Passes, Recitals and Makeup Classes
A practical guide to an AI agent for dance studio pass balances, recital logistics and makeup requests, with clear credit rules and human approvals.

An AI agent can handle the questions a dance studio front desk answers most often: how many classes remain on a pass, whether a missed class can be made up, and what the next step is for a recital. It works only when the studio has written its credit rules as plain facts and the agent reads them from one ledger instead of guessing. This guide explains how an owner or operations lead can set that up, what the agent should decide alone, and where staff must keep the final say.
Key takeaways
- Write the credit rules as short statements before any agent touches a balance: what a credit buys, when it counts as used, and when it expires.
- Let the agent answer pass and schedule questions from the ledger of record, never from memory of an earlier conversation.
- Treat a makeup booking as a request with a rule check and a single reference, so retries cannot consume a second credit.
- Keep recital logistics in one plan with a named owner for each item, so fees, costumes and rehearsals stop arriving in separate threads.
- Review the agent's answers and its permissions on a schedule, because studio rules change with each season.
Why pass balances and makeups fill the front desk
Most studio messages about money and attendance are small, repetitive and time sensitive. A parent asks how many classes are left before a trip. A student wants to know whether a missed class can be replaced this week. A family asks whether the recital fee already covers costume and shoes. Each question takes a few minutes, but together they interrupt teaching and front-desk work throughout the day.
The root problem is rarely the software. Credits are often spread across a booking system, a payment record, a spreadsheet and a teacher's notes. When those sources disagree, staff spend time reconciling them, and students lose trust in every answer. An agent connected to all of them without a rule about which source wins makes the problem worse. Choose one ledger of record for passes and attendance, then let the agent answer from it. If your studio also runs memberships and waitlists, the yoga studio guide covers a similar pattern from the membership side.
Write the credit rules before the agent sees them
An agent cannot apply a rule that nobody has written down. Before you configure anything, state the rules in terms a new front-desk hire could follow. Settle these points first:
- What one credit buys: one group class, one private lesson, or one drop-in at a specific level.
- When a credit is used: at booking, at check-in, or when the class starts. Pick one and state it.
- Expiry and renewal: whether unused credits end with the pass period and whether a renewal carries anything over.
- Cancellation window: the notice you require, and what happens to the credit inside and outside that window.
- Makeup eligibility: which absences qualify, how long a student has to use the makeup, and which classes accept makeups.
- Exceptions: who may grant one, under which circumstances, or nobody.
Write these on a single page and date the version. When a rule changes, you will know exactly which version the agent was given. A rule that lives only in someone's head is the thing that produces most disputes.
Makeup classes without double-spending credits
The riskiest action in this workflow is the one that consumes a credit. A student who taps a link twice, or an agent that retries after a timeout, should not end up with two bookings and two debits. Payment platforms handle this with idempotency: the caller sends a unique key with a request, and a repeated request carrying the same key returns the original result instead of acting again. The Stripe documentation on idempotent requests describes the pattern clearly, and a studio can apply the same logic to any booking or credit change.
A reasonable makeup flow follows these steps:
- Confirm the absence is recorded and falls inside the makeup window in your rules.
- Read the pass ledger and confirm an unconsumed credit exists in the current pass period.
- Check seat capacity in the target class, along with any level or age restriction.
- Create one pending hold under a request reference, so a retry returns the same hold.
- Send the options in writing, and debit the credit only after the student confirms and the seat is secured.
- Release the hold and restore the credit automatically if the student cancels inside your rule.
Illustrative scenario: a student misses a Tuesday class and writes on Wednesday evening asking for a makeup. The agent checks the ledger, finds one credit left in the current pass period, and sees an open seat in a Thursday class at the same level. It offers the Thursday seat and a Saturday option. The student picks Thursday and confirms, so the agent creates one booking under one reference and debits one credit. If the student sends the same request again that night, the agent returns the existing confirmation instead of creating a new booking.
Recital logistics in one plan, not many threads
A recital touches registration, costume sizing, fees, rehearsal calendars, ticket requests and volunteer roles, and many of these questions come from the same families. The useful pattern is one recital plan with a named owner for each item. The agent answers questions from that plan, tracks which families have completed each item, and drafts reminders that a staff member approves before they go out. It should not promise a refund, a costume exception or a change to a child's role in the performance.
Use this decision guide when a recital question arrives:
- Is the answer already in the recital plan? The agent answers and points to the plan entry it used.
- Does the question change money, such as a fee waiver or a refund? Staff approve before anything is promised.
- Does it change a child's role, placement or costume? The teacher decides.
- Is it a complaint, an injury or a safety concern? A person responds directly, and the agent only acknowledges receipt.
The result is fewer separate threads and one record of what was promised. Studios that also keep a waitlist for popular recital slots or intro classes can apply the same discipline. The climbing gym guide shows how waitlists can run under clear rules.
Where the agent acts alone and where staff decide
The agent should hold the permissions the job needs and no more. It can read the ledger and the recital plan, create a pending hold and draft a message. It should not edit a balance directly, approve a refund or change a pass end date. Those permissions belong to the system and to named staff, not to the prompt. The OWASP authorization guidance makes the general point: access decisions should be enforced by the application, not left to the model's instructions.
Parents also paste text into messages that reads like instructions, sometimes by accident. The agent should treat anything a customer writes as content to answer, never as policy that overrides your rules. Write the boundary as a short list the owner can read in a minute: what the agent may do, what it may prepare for approval, and what it must never touch. Review that list whenever a staff member joins or leaves, because permissions tend to accumulate over time.
How Be Human runs this as a managed service
Be Human starts with one useful role, here a front-desk agent for pass balances, makeup requests and recital questions. The first engagement defines the memory the agent may use, such as the credit rules, the class catalog and the recital plan. It also defines the tools it can call, such as ledger reads, seat holds and a staff approval queue, along with the approval boundaries and one observable success criterion, for example balance answers that match the ledger in a regular review sample. Managed operations then cover monitoring, recurring evaluations on real conversations, permission review when staff change, integration maintenance when the booking system updates, and incident response if the agent gives a wrong balance. The service is remote and available in English and Portuguese.
Choosing the first studio task
Most studios should not start with everything at once. Begin with the question that generates the most repeated messages and the fewest exceptions, which is often the balance question. Add makeup requests once the ledger is reliable, and bring in recital logistics after the plan is written and owned by named people. If you are not sure which front-desk task is the right first one, take the free Be Human AI diagnostic to find the first task worth handing to an agent.
Frequently asked questions
Can an AI agent tell a student how many classes are left on a pass? Yes, if it reads the balance from the ledger of record at the moment of the question and the studio has defined when a credit counts as used. The answer should include the pass period and any pending holds, so the student does not receive a number that changes after the message is sent. If the ledger is incomplete, the agent should say it cannot confirm and route the question to staff.
Should the agent approve makeup classes on its own? Only within rules you have written and only for standard cases. The agent can check eligibility, hold a seat and prepare an offer, but a staff member should approve exceptions such as makeups outside the window, transfers between students or waived cancellations. Keep the approval step visible in the workflow so you can see who decided and when.
How do you stop an agent from using two credits for one makeup? Give every booking action a unique request reference, and make the system return the original result when the same request arrives again. Debit the credit only when the booking is confirmed, and release it automatically when the student cancels inside the rule. Test a deliberately repeated request before you rely on the workflow.
What should a studio measure after launch? Measure whether balance and schedule answers match the ledger in a regular review sample, how many messages are escalated and why, and how often staff correct a draft before it is sent. Review rules and permissions when the season changes, not only when something goes wrong.
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.


