AI for Pilates Studio Scheduling: Classes and Makeup Rules

AI for Pilates Studio Scheduling: Classes and Makeup Rules

How an AI agent can manage class scheduling and makeup classes at a pilates studio, with written rules, clear limits and handoff to the instructor.

•

Want to apply this in your company?

Book a call with Juan →

An AI agent can run the schedule of a pilates studio: check open spots by class, record attendance and no-shows, apply the makeup rule the studio already follows, and confirm everything over WhatsApp. It works well when the rule is written down, the timetable is up to date, and any question about health or how to perform an exercise goes to the instructor. Physical assessments, pain, injuries and contraindications should not be handled by the agent.

Key takeaways

  • Scheduling, spot availability and makeup classes are rule-based processes, well suited to automation with clear limits.
  • The makeup rule has to be written down before the agent is involved: notice period, accepted reasons, who can make up a class and until when.
  • Questions about health, pain, injury or whether an exercise is suitable belong to the instructor, and the agent only routes them.
  • Student data is personal data, and health-related data calls for extra care under Brazil's data protection law (LGPD).
  • Start with one class and one flow, log every decision, and only then expand.

Why pilates scheduling is a good first workflow

A studio's operation usually revolves around recurring classes, limited spots and last-minute cancellations. In many studios, control still lives in a spreadsheet, a notebook and the personal WhatsApp of whoever is answering messages, so the information depends on one person's memory. This is common, and it is exactly where an agent helps: it replies at any hour, always applies the same rule, and keeps a record of every conversation.

The real gain is not a faster reply. It is a consistent one. When a student asks whether they can make up Thursday's class, the answer should follow the same rule for everyone, not depend on who was on duty. This reduces friction with the team and avoids promises the studio cannot keep.

What the agent should do and what stays with the instructor

The division of responsibilities is the most important part of the project. Before configuring any tool, write the two lists below with your team.

  • For the agent: share the timetable and open spots for each class, book and cancel classes according to the rule, log a no-show with the date and time of the notice, offer a spot in a makeup class, and place the student on a waiting list.
  • For the agent: route to a human any request for an exception, complaint, disputed payment, and any message that mentions pain, injury, pregnancy, surgery or medication.
  • For the instructor: decide whether a student can return to training after an injury, adjust exercises, assess posture and define progression. The agent does not suggest exercises or interpret symptoms.
  • For management: define exceptions, approve makeups outside the rule and review routed conversations on a regular basis.

If the team cannot state in one sentence what the agent never answers, it is not yet time to put it on WhatsApp.

How to define the makeup rule before automating

An agent only executes well what is written down. Many conflicts with students come from rules that exist only in the head of whoever answers the phone. Before automating, document the points below.

  • The minimum notice required to cancel without losing the class, in hours or days, and exactly when that period starts counting.
  • The reasons that entitle a student to a makeup, such as a medical certificate, a trip communicated in advance or a studio holiday, and the reasons that do not.
  • How many makeups are allowed per period and how long each credit stays valid.
  • Which classes are eligible for makeups, considering level, format and available spots.
  • What happens to a class that is not made up: loss of the credit, extension or another rule the studio sets.
  • Who can approve exceptions and in which channel that decision is recorded.

Note that the studio sets the commercial and operational rule. The agent only applies what has been decided. If a rule depends on case-by-case judgment, it should be escalated to management, not decided by the model.

A seven-step plan for a scheduling pilot

A pilot should be small, measured and reversible. The sequence below works for one format or one recurring class.

  • Step 1: map the current process, including how a no-show reaches the studio today and how long it takes for the freed spot to be offered again.
  • Step 2: write the makeup rule and the list of agent limits, validated by management and the responsible instructor.
  • Step 3: connect the agent only to the calendar of the chosen class, with read and booking permissions, and no access to financial records.
  • Step 4: test with simulated conversations before any real student is involved, including late notices, waiting lists and messages about health.
  • Step 5: activate the agent for that class, with visible human handoff and a record of every decision.
  • Step 6: agree on a success metric before starting, for example the share of makeups resolved without manual intervention, and review routed conversations every week.
  • Step 7: decide whether to expand, adjust or stop. If you expand, document the rule and hand the operation to the person who will maintain it.

Illustrative scenario: a no-show with a late notice

Illustrative scenario: a student messages on WhatsApp saying she will not make the 6 pm class and asks whether she can make it up. The agent checks the registered rule, sees that the notice arrived after the minimum period and that the reason she gave is not on the list of exceptions. It replies with the rule applied, lists the makeup spots available within the credit's validity, and offers a place on the waiting list for another class.

If the same student says she has pain after the last session, the agent does not assess the situation or suggest that she keep training or stop. It tells her the conversation will be passed to the team and logs the request with the time. The responsible instructor receives a summary of the history and makes contact. In this case, the schedule is only the background, and the clinical decision stays with the person qualified to make it.

How to handle student data under the LGPD

Student data includes name, phone number, attendance history and, in some cases, health information. Under the LGPD, health data is classified as sensitive personal data, which requires additional care. For scheduling, the agent does not need this data. Configure the flow so the conversation does not collect clinical information, and ask students not to send medical details through the chat.

A few practical steps help. Limit the agent's access to what it needs to book and cancel classes. Define how long conversations are stored and who can review them. Tell students they are talking to an automated system and how to reach a human. For official guidance on the law and on personal data processing, check the materials published by the Brazilian National Data Protection Authority (ANPD).

It is also worth checking the WhatsApp customer service window before sending reminders or makeup offers, because sending rules change depending on the type of message. This is covered in WhatsApp 24 Hour Customer Service Window: How an AI Agent Should Act.

How we apply the method in a scheduling pilot

At Juan Carlo Gonçalves, the work follows four stages: map the current process and its baseline, limit data and actions, pilot a real flow with records and an agreed metric, and then hand over the operation and decide whether to expand, adjust or stop. For a studio, this means starting with the written makeup rule, not with the tool. A scheduling agent that cannot route a pain report to the instructor is not ready, no matter how well it is configured. Some scheduling needs are similar in other service businesses with recurring appointments, such as a booking agent for barbershops without empty slots, and the logic of limits tends to be alike.

If your studio wants to evaluate this, the first step is a diagnostic conversation about the current flow, with no commitment to implementation.

Frequently asked questions

Can the agent answer questions about exercises and injuries? No. Questions about pain, injury, contraindications or whether an exercise is suitable must be routed to the instructor or the responsible team. The agent can tell the student the conversation will be transferred and log the request with the time and a summary. This separation should be written into the instructions and tested before activation.

Can the agent apply the makeup rule automatically in every case? Only in the cases the written rule covers. Situations outside it, such as exceptions, complaints or special requests, should go to management. The clearer the rule, the fewer cases reach the team, but some will always require a human decision.

Do we need to replace the studio's scheduling system to use an agent? Not always. In many cases the agent can be connected to the existing calendar with restricted permissions, provided the integration is possible and secure. The answer depends on your current tool and should be decided after mapping the process.

How will we know if the pilot worked? Before starting, define a simple metric and a review date, such as the share of makeups resolved without manual intervention and the number of routed conversations by reason. Review the records with the instructor and management. If the result is not what was agreed, adjust the rule or stop the pilot without expanding it.

Sources

pilates studioAI schedulingWhatsApp automationLGPDAI agents