Using inTusell at a gym: a daily operations guide

In the core article you trained inTusell’s AI. You uploaded your membership packages, your class schedule, your price and campaign notes, and your frequently asked membership questions to the knowledge base, marked your sales conversations, and set the persona and gym-specific boundaries. Training was a one-time setup. Now the real work begins: how does a gym / fitness / pilates studio use inTusell: that is, how does gym inTusell usage work in daily operations? This article shows, step by step, what the assistant you trained does in the field across a workday: from the "how much is membership?" question arriving in Instagram DM to a trial class, and from class booking to member follow-up and handover to a human.

This article is the second part of the gym section of our sector-based "how to train your AI" series. The first part, how you train inTusell’s AI, explained the setup; this article explains turning that setup into daily operations.

Quick answer

The flow of a day

The typical day of a studio that has set up inTusell looks very different from before. The front desk or the owner used to arrive in the morning and go through dozens of unread Instagram DMs and unanswered WhatsApp messages one by one, sorting which one was asking about price, which about a class time, which about a membership freeze. Now the nighttime messages are already answered, requests are pre-classified, some trial classes are arranged, and reminders have gone out. The backbone of a day looks like this:

The team no longer spends its day scanning messages and sorting "is it price, a class time, or a freeze." It works on what truly needs a human touch: showing the gym, negotiating with a prospect, welcoming a member.

Who is it for?

This usage model is especially meaningful for these sports businesses:

• Studios that receive heavy traffic from Instagram and WhatsApp: those who can’t keep up with dozens of "how much is the price, is there a trial?" DMs a day.
• Gyms that lose the prospect after a trial class: those with a flow of prospects who come to a trial, don’t return, and go to another gym because no one followed up.
• Businesses that miss after-hours and evening demand: those for whom, by the time they see the nighttime interest in the morning, the prospect has already signed up elsewhere.
• Pilates/fitness studios that run the class schedule, membership, and social media together with a small team: those forced to track the front desk, DMs, and class capacity separately.

If your message volume is more than one person can keep up with, inTusell meets it like an experienced front-office assistant with clear boundaries, not like a chatbot.

The most important boundary: the AI does not give medical and injury advice

This is the foremost rule for a gym. inTusell does not give a program recommendation in a concrete health situation, a "do/don’t do this exercise" type of direction, or treatment advice. That is a professional boundary, and it also keeps a wrong steer from injuring the member. The binding assessment is made by the trainer, the instructor, or, when needed, a physician.

In practice, what the AI does and doesn’t do splits like this:

When a member asks a concrete question like "I have knee pain, which classes can I do?", the AI doesn’t list programs. It says "Let’s assess this with our instructor; it would also be good for you to consult your physician if needed" and routes them to the instructor. Wherever it doesn’t know, or wherever expertise is required, it states its boundary honestly and hands over. We explain how to set this boundary during training in how you train inTusell’s AI.

Incoming request: the member asks, the AI gives the information and routes

The flow begins when a member sends a message. Imagine someone writing in Instagram DM "How much is pilates membership, can I do a trial first?". The AI reads the message and pre-classifies the request: here, both a price query and a trial request. It then explains the membership packages from the knowledge base and directs the prospect to a trial class or an intro meeting.

Two points are critical. First: the AI does not state price and campaign out of thin air. It shares only the current package and campaign information you uploaded to your knowledge base. Second: whatever language the member writes in, the AI answers in that language: a practical convenience for boutique studios with foreign members.

So the prospect reaches the team complete, and nobody has to question the member again. You’ll find the detail of the membership and class booking engine in membership and class automation; and we explain how all channels merge in a single inbox in Instagram and WhatsApp automation.

Membership and package questions: answers from the knowledge base, no making things up

In a gym, most incoming messages are not a new person but the same recurring question: "How much is monthly?", "Is there a student discount?", "Is there a freeze option?", "How many times a week is the reformer class?", "Is there a locker room and shower?" These questions cycle in the same pattern all day and eat up the front desk’s time. inTusell takes this load by answering from the knowledge base.

The most critical rule here is this: the AI does not make up price and package contents. It gives the answer only from the source you uploaded to your knowledge base:

If someone asks about information that isn’t in the knowledge base (a campaign you didn’t upload, say, or a personal discount) the AI doesn’t make anything up. It says it doesn’t know and offers to connect the member to the team. That keeps you from giving a member a wrong price and then having to say "actually that campaign ended." A wrong price damages both the member relationship and the gym’s credibility, which is why the "say you don’t know" behavior is a protection, not a constraint.

How the knowledge base is fed and stays current

Whether the AI gives the right answer depends entirely on how current your knowledge base is. inTusell feeds the knowledge base from several sources and indexes them all for pgvector-based search (RAG):

• Package and price files: uploaded as PDF, Excel, or CSV; membership types, package contents, price ranges, and freeze terms come from here.
• Web URL: you provide the address of your studio page, your class schedule page, or your FAQ page; the content is fetched, indexed and re-scanned weekly.
• Instagram profile: you import your last 100 posts, so your class and campaign announcements enter the knowledge base.
• FAQ suggestions: suggestions come from the questions members ask most; the ones you approve are added to the knowledge base.

A short campaign note like "2 months free on annual membership this month" also goes into a file or onto your website and from there into the knowledge base; the panel has no separate free-text box.

In a gym, campaigns and the class schedule change often. So when you update your prices or your schedule, refresh the knowledge base too. The AI always speaks according to the last source you uploaded, and if the source isn’t current, neither is the AI. That is why keeping prices and the schedule current is the backbone of operations.

An additional security layer: sector-specific instructions constrain the AI from the start so that it isn’t inclined to make things up. On top of that, an output filter (a pattern scan plus a second-model check) scans responses and flags misleading health or result promises (for example "you’ll definitely get results," "you’ll lose weight with this program"). These two layers reduce the risk but are not a guarantee; the real safeguard is a current knowledge base.

Trial class and class booking: slot, capacity, and conflict prevention

When a member requests a trial class or a group-class registration, the AI doesn’t give a time out of thin air. For one-to-one appointment types (for example "Trial class 50 min") it looks at the duration, your working hours and whether that time is booked. For group classes you define the weekly program under Appointments › Settings › "Group Sessions" ("Saturday 10:00 reformer, capacity 8", say); the AI offers those sessions with their remaining seats and enrolls the member in the one they pick. If you haven’t defined group sessions, the AI doesn’t invent any. This is the appointment engine’s most visible job in a gym: "can I sign up for the reformer class at 19:00 tomorrow evening?" is solved with a single message.

The most critical protection here is the capacity check: no one can be enrolled in a full session. The capacity check and the enrollment happen in a single database transaction, so two members cannot take the last seat at the same time. For one-to-one appointments, a second booking at the same time is not opened. A cancelled booking frees its seat again. Every booking gets an automatic reminder 1 day and 2 hours before; the 1-day message asks for confirmation ("1 Yes · 2 Reschedule · 3 Cancel") and carries a change/cancel link. A member who can’t come gives the seat back early, and fewer members fail to show up. Reminders run on paid plans. One-to-one appointments are copied one-way to Google Calendar (inTusell → Calendar); calendar sync does not run for group-session bookings, whose occupancy and attendee list appear on the appointments screen. This module can be enabled on every plan; if you don’t need it, you simply don’t use it.

Member follow-up and conversion: the follow-up engine’s most valuable job

In a gym, the prospect you lose most often is the one who comes to a trial class, likes it, says "let me think," and never returns: forgotten because no one followed up. Winning over someone who already came to a trial is usually easier than finding a brand-new prospect, which makes this the follow-up engine’s most valuable job in the gym. inTusell uses a proactive follow-up flow to track prospects who took information or a trial class and then went silent.

The AI writes the follow-up messages itself (they are not templates), sends at most 3, and only within business hours (your appointment hours, otherwise 09:00-21:00 local time) in the prospect’s WhatsApp, Instagram or Telegram conversation; web chat gets no follow-up. It stops when the prospect replies, books, or says "don’t write to me". The message isn’t a dry "come to us" call; it directs the prospect either straight into the membership flow or to a second trial/intro meeting. For an interested prospect, the AI collects the needed current information (for example which package they looked at, which classes they’re interested in) and hands the prospect over to the team fully prepared.

One boundary should be clear: inTusell is not a separate list product for prospects who went quiet. What it does is send a well-timed follow-up message to a prospect who talked to you and went quiet, for example a message to someone who took a trial class and didn’t return, along the lines of "How was your reformer trial last week, shall we continue with a membership?" When you sit down at operations in the morning, you can see which prospect is warm and which is awaiting a return, so you no longer keep a manual reminder list and work down it call by call. We explain in detail how this engine is set up and how it merges with the membership flow in membership and class automation.

Health and injury questions: inform, then route to the specialist

A member often writes out of health concern: "I have a herniated disc, is pilates suitable for me?", "I had surgery, when can I start?" inTusell behaves with balance at this moment. It doesn’t give medical advice, doesn’t rule "you can/can’t"; it shares the general class contents and suitability notes you entered into your knowledge base, routes the situation to the instructor/trainer, and says to consult a physician when needed.

So on health matters the AI is an informing and routing layer: it carries the member to the right person without misdirecting them. Determining the program, assessing fitness, and any adaptation in case of injury always stay with the specialist. The member feels taken seriously on a sensitive matter, and the instructor picks the situation up with the context the AI collected, not from scratch. As in law and health, inTusell here gives information, not advice.

Following up the silent prospect: proactive follow-up and CRM

In a gym, most sales are lost to the prospect who says "let me compare prices" and disappears, or who keeps postponing the trial. inTusell doesn’t forget these prospects. Up to 3 follow-up messages go to the person who took information or a trial and then went silent, so the prospect isn’t forgotten. Every conversation is recorded in the CRM, so the prospect’s history, which package they were interested in, which classes they looked at, and which stage they stopped at all stay on record. The next contact starts from where they left off, not from scratch.

One boundary should be clear here too: inTusell does not keep a managed "waiting list." What it does is send a well-timed follow-up message to a prospect whose conversation stalled, for example a reminder to a prospect who looked at the annual package and went silent, along the lines of "The campaign on the annual membership you asked about is still on, shall we look at it?" The AI mentions a campaign only if it is written in your knowledge base. When the team sits down at operations in the morning, they can see which prospect is warm and which is awaiting follow-up. There’s no need to keep a manual reminder list.

The gym pipeline

Every prospect moves along a pipeline, and the AI advances them to the right stage as the conversation develops. These are the stages a gym setup comes with:

New Request
  → Information Given
  → Appointment Booked
  → Service Done
  → Sale Completed (won)
or → Dropped Out (lost)

A conversation counts as won only when a real record such as an order or an appointment exists, not on the AI’s own guess.

The pipeline shows you at a glance which prospect only asked about price, which came to a trial class, and which is awaiting a membership decision. A follow-up chain runs for the prospect who goes quiet after a trial, so new interest and post-trial follow-up move through the same panel without getting mixed up.

Handover to a human: operating modes

You decide how much the AI does on its own. There are three account-wide modes (My business › Instructions › AI Settings › Advanced AI Settings):

In every mode, the AI hands a medical or injury question, a personal negotiation request, or a question outside the knowledge base to the team. The handoff lands on the Decisions screen with the full conversation history, so the member doesn’t have to start over. If you want one conversation handled differently, for example a corporate group agreement, you switch that conversation to Approval, Shadow or Paused.

An additional note: a file such as an old membership document or a campaign image is not on its own a reason for handover to a human. The AI takes the attachment, understands the context, and continues the flow; it hands over only if there’s a situation that truly requires a decision (for example an active health issue, a negotiation).

Privacy, KVKK, and sensitive data

In a gym, incoming information is often sensitive: name, phone, and in some cases a health declaration and injury history. inTusell gives you these tools:

1. Consent record: Consent is recorded per channel and purpose, with its time; the member can see and change it from the consent portal link.
2. Legal notices: "Legal notices" under My business › Business Information sends the member a short KVKK notice with your own privacy link at first contact and adds an opt-out line to promotional messages. It is on by default; only the owner or an admin can change it.
3. Sensitive field control: The AI never extracts high-sensitivity fields such as a health restriction disclosure from the chat on its own and never saves them without your confirmation.
4. Encrypted storage, separate data and deletion: Data is stored encrypted and each business’s data is kept separate. A retention policy deletes expired data nightly; you export with "Download My Data (KVKK)". In the sales call recordings you upload, phone numbers, emails, ID and card numbers are masked in the transcript.

This makes it easier to answer, with proof, the "what did the member accept and when?" question in an audit or dispute; at the same time it prevents sensitive information from being stored carelessly.

Reports: what’s working, what to fix?

To make operations visible, you track basic metrics in the panel: how many messages came from which channel, how many the AI answered, how many went to the team, how many people sit at which stage of the sales pipeline, and how many appointments and class bookings were made. Seeing these numbers serves two purposes:

1. Finding knowledge-base gaps: If there’s a package or class the AI often hands over saying "I don’t know," that information is missing from the knowledge base. Once you add it, the AI answers that question itself.
2. Improving response rules: When you see a wrong or missing answer in Conversations or in AI Test, you add the fact to the knowledge base or write a rule in AI Rules ("ask which class they’re interested in before giving a price," for example). Corrections in the Label Manager are a quality record; they don’t change how the AI writes replies on their own. You can also compare two tones (formal and friendly, say) with an A/B test (beta).

So reports are not just a summary; they are the feedback loop that sharpens the assistant over time.

What it isn’t

To put inTusell in the right category, let’s clarify what it is not:

• It is not a trainer or a health advisor. It doesn’t draw up a program, doesn’t assess injury, doesn’t give medical advice; the instructor, trainer, and, when needed, a physician handle these.
• It is not a membership/POS software. It doesn’t open the membership automatically, doesn’t collect payment; it gathers the membership request and carries it to the team: the gym opens the membership.
• It is not a class-studio management dashboard. It handles capacity and occupancy control with the appointment engine, but it does not handle business management like instructor shifts or payroll; it’s a layer for communication and coordination.
• It is not a gym-specific priced product. Plans differ only by messages, voice minutes and storage, not by sector; channel count is never charged. The appointment module can be enabled on every plan; follow-up and reminders run on paid plans.
• It is not a simple chatbot. inTusell is an AI sales assistant and CRM that answers messages on WhatsApp, Instagram and web chat, books appointments and opens orders, and follows each customer through the sales pipeline.

One more note: Instagram comment replies are live and on by default, but if you connected Instagram before September 24, 2026, you need to reconnect it once; until then no comment gets an AI reply. Keeping these boundaries clear from the start saves you from setting the wrong expectation. For product and price details see the solutions and pricing pages.

Frequently asked questions

Does the AI sell memberships on its own?

The AI answers membership and package questions from the knowledge base and carries a suitable prospect all the way to a trial class or an intro meeting. The contract, payment, and opening the membership stay with you. In situations outside the knowledge base or requiring negotiation, the AI hands the conversation over to the team. The gym always opens the membership.

Does the AI make up price and package information?

No. The AI states the membership fee, package contents, and campaign information only from the source you upload to your knowledge base. It does not make up a package, discount, or condition you didn’t upload; it says it doesn’t know and hands over to the team. This is the most critical trust rule in a gym.

How does it handle trial-class and class booking?

If the trial class is a one-to-one appointment type, the AI checks your working hours and booked times, offers a free time and books it. If you defined group classes in the panel as group sessions with capacity, it offers the sessions that still have seats and enrolls the member; it never books into a full session. Every booking gets a reminder 1 day and 2 hours before, so fewer members fail to show up.

Can a member be connected to a representative?

Yes. The AI hands a conversation it is unsure about, or one that needs a person, to your team; the handoff lands on the Decisions page with the full chat history. Account-wide you choose Intern (default), Senior or Full approval; you can also switch a single conversation to Approval, Shadow or Paused.

What does the AI do with a health or injury question?

The AI does not give medical advice. For a question like "Is this exercise suitable for my herniated disc?" it doesn’t recommend a program; it directs the situation to the trainer or instructor and, if needed, says to consult a physician. Here the AI is a layer for informing and routing to the right person; the specialist determines the program.

How are member data and KVKK protected?

Data is stored encrypted and each business’s data is kept separate. Consent is recorded per channel and purpose, with its time. The AI never saves high-sensitivity fields such as a health restriction disclosure without your confirmation. With "Legal notices" on, the member gets a short KVKK notice with your own privacy link at first contact.

Next step

If you haven’t trained your assistant yet, start with how you train inTusell’s AI: the foundation of this operation is built there. To go deeper on the channel side, move on to Instagram and WhatsApp automation. For the details of the membership-collection and class-booking engine, see membership and class automation. The gym-specific solution is on the gym page, and the whole series is on the all posts page.

If you wonder how the same logic (with its "doesn’t give advice, routes to the specialist" boundary) is set up in another field, clinic AI training offers a good comparison; and in tourism the flow is in tour agency AI training.

To see live how it would work at your own gym, use Get a demo or write directly to hi@intusell.com. In a short session we’ll open your inbox together and test a first membership question and a trial-class booking in the system.