Treat the intro offer as a guided sequence rather than a discount: get the first three classes booked at purchase, send messages triggered by attendance (after class one, a nudge if nothing is booked within five days, a membership conversation before the offer expires), and use AI to personalise each from instructor notes. Then measure conversion by monthly intro cohort.
AI's contribution is personal messages at scale, written from what the instructor noticed and timed to what the client actually did. It can't rescue a badly designed offer, so the offer comes first. The work splits three ways: diagnosis (where your intro clients drop out), design (the offer and the class-by-class journey) and delivery (instructor notes, the membership conversation and your software's automations).
Find where your intro clients drop out
Before writing a single message, count the last three months of intro buyers through five stages. Most studio software can report this; if not, export bookings and ask ChatGPT or Claude to build the table by purchase month.
| Stage | What a big drop here usually means |
|---|---|
| Bought the intro offer | Starting point |
| Attended class 1 | Booking friction, nerves, or a gap of too many days between purchase and first class |
| Attended class 2 | The first class experience: too hard, too full, nobody said hello |
| Used the whole offer | Timetable fit: the classes they want are full or at the wrong times |
| Bought a membership within 14 days of expiry | The price step, the ask itself, or no clear recommendation |
Check the table before trusting it. A common slip when a chatbot builds it from an export: the product column holds both "Intro 3 classes" and a regular "3-class pack", and the model counts both as intro buyers. The first sign is a total that looks too high (say 71 intro buyers in a month when you know you sell about 45) and a class-1 attendance rate that looks suspiciously good, because returning clients buying a pack always turn up. Ask the model to list the distinct product names it counted and confirm which ones are intro offers, then have it rebuild the table from those alone. It takes two minutes and saves you designing a fix for a drop-off that isn't there.
Each drop has a different fix, and only some of them are messages. If most people vanish between class 1 and class 2, a better after-class message helps, but so does making sure a new client never lands in a packed advanced class. If the drop is at the end, the problem is the conversation about membership. And if the classes intro clients want are always full, using booking data and AI to fix your class timetable matters more than any sequence.
Design the offer so it can convert
Intro offers come in a few shapes, each with trade-offs:
| Offer shape | Strength | Weakness |
|---|---|---|
| 3 classes in 14 days | Cheap to deliver, a clear finish line | Too short if the client has one busy week |
| 2 weeks unlimited | Builds a habit fast | Attracts people collecting free trials |
| Intro private session plus 2 group classes | Safe start on the reformer, a relationship with one instructor | Costs instructor time |
| First month at a reduced membership rate | No step from "intro" to "member"; they already are one | A bigger commitment for a nervous beginner |
Two design choices matter most for conversion. First, the expiry: long enough that a client with one disrupted week can still use all their classes. Second, the price step: if membership looks expensive next to the intro price, crediting the intro fee against the first month for anyone who joins before expiry turns the step into a reward for deciding. Whichever you pick, write it down plainly; the AI messages will quote it.
A quick sum shows why the credit works, using illustrative prices. The intro offer is 3 classes for $49, drop-in is $32 a class, and the twice-a-week membership is $180 a month. Seen side by side, $49 to $180 feels like a jump of more than three times. Credit the $49 against the first month for anyone who joins before expiry and the first month costs $131; at eight or nine classes a month, the ongoing membership works out at about $20-$23 a class against $32 drop-in. Your AI messages should quote those two figures, the first-month cost and the per-class saving, rather than the headline membership price on its own.
Beginners and experienced movers need different journeys
Intro buyers split into two groups who want opposite things. Complete beginners want reassurance: what happens in the first class, whether they'll keep up, which class is gentle enough. People who've done Pilates elsewhere want to know whether your classes are challenging enough and which instructor to try. Ask one question at purchase, "Have you done reformer Pilates before?", and let the answer steer everything after it.
- Beginners: book them into foundations or beginner-labelled classes, keep the first messages about comfort and what to expect, and introduce the instructor by name.
- Experienced: point them to the right level straight away, mention the class formats that differ from other studios, and skip the basics.
Pass the answer to the AI in every prompt ("client is new to reformer" or "client has done reformer for two years elsewhere") so the after-class messages and class suggestions match. It's a small change that stops experienced clients being sent to beginner classes they'll find dull.
The intro journey, class by class
Trigger messages from what the client does, not from the purchase date alone. A client who attended twice in four days needs a different message on day five from one who hasn't come at all.
| Moment | Message | AI's part |
|---|---|---|
| Purchase | Welcome, what to wear and bring, arrive ten minutes early, and links to book classes 2 and 3 now | Suggests beginner classes at times close to the first booking |
| Evening before class 1 | Reminder, and "tell your instructor about any injuries before you start" | None needed |
| After class 1 | How did it feel, plus one personal line from the instructor | Drafts from the instructor's note |
| Day 5, no second booking | Two specific classes that fit the times they chose before | Picks the classes |
| After the last intro class | What they've done so far and a membership recommendation | Works out the fit from attendance |
| 3 days before expiry | A reminder of any intro credit and the recommendation | Drafts; owner or instructor sends |
| 7 days after expiry, not joined | One question: what held you back? | Sorts the replies |
The day-5 nudge is where automated class picks most often go wrong. An illustrative first run: a beginner who came to Tuesday 7pm Foundations got "Why not try Thursday 7pm Reformer Flow or Saturday 9am Reformer Burn?" Both fitted her times. Both were intermediate classes, and Saturday's had no spaces left. The fix is in what you feed the model, not the wording: pass only classes with at least two free spaces, filtered to the level from the purchase question, for the next seven days. With that list in the prompt, the same client gets Thursday 7pm Foundations and Sunday 10am Foundations, which is what an instructor would have suggested.
The post-expiry question produces short, blunt replies, and the point of sorting them is to do something different with each kind:
| Reply (illustrative) | Category | What happens next |
|---|---|---|
| "Loved it but can't stretch to $180 right now" | Price | One reply from the owner mentioning the class pack; no discount |
| "The evening classes were always full" | Timetable | Logged for the timetable review; offer a waitlist spot |
| "Just wanted to try it, thanks" | Trying it | A short thank-you; no further contact |
| "My back was sore after the second class" | Health | Instructor replies personally; nothing automated |
| "Moving house next month, will be back" | Timing | A "don't contact until" date, then one message |
After a few months, count the categories. If "timetable" outnumbers "price", the next improvement is a class, not a message.
Turning instructor notes into personal messages
The after-class message is where a studio feels different from a gym chain, and it's the one that's hardest to keep up by hand. The fix is a one-line note from the instructor after each class for every intro client (in the client notes of your studio software or a shared sheet), then AI to turn it into a message, and the instructor approves the batch before leaving the studio.
Turn each instructor note into a short message from the instructor
to the client, 2-3 sentences, in the instructor's voice.
Instructor voice sample: [paste 2 real messages they've written].
Note format: client first name | class | instructor note.
Rules:
- Only use what's in the note. Add no details, compliments or facts.
- No comments about the client's body, weight or "weak" areas.
- If the note mentions pain, injury or pregnancy, write nothing and
output "INSTRUCTOR TO REPLY PERSONALLY".
- End with the next class suggestion if the note includes one.
Notes:
[first name] | Reformer Foundations Tue 7pm | found footwork tricky, laughed
through it, loved the stretch at the end; suggest Thu 7pm FoundationsWhich might become: "Hi [first name], lovely to meet you on Tuesday. Footwork takes everyone a few classes, and you handled it with a smile. If you liked that final stretch, Thursday's 7pm Foundations ends the same way. Hope to see you there. [instructor]"
The rule "add no details" matters, because models like to embellish. Here's how it shows up. The note says "quiet, stayed after to ask about springs". The draft says: "You were a natural on the reformer and clearly have a strong core already. Loved your questions about the springs!" Only the last sentence comes from the note. "A natural" and "a strong core" are invented, and the second is a comment on the client's body, which the rules also ban. The instructor's version: "Thanks for staying on to ask about the springs. Ask me any time, it's the part everyone finds confusing at first." Five minutes of reading before the instructor goes home catches it. For keeping these messages sounding like your instructors rather than a template, using AI in a Pilates studio while keeping your voice goes further.
The membership conversation
A recommendation based on what the client actually did is easier to accept than a price list. Match the plan to their attendance:
- Three classes in ten days, and asking about more: the twice-a-week or unlimited membership.
- A steady once a week: a four-class monthly pack or a once-a-week membership.
- Spread out and irregular: a class pack, with no pressure to commit.
AI is handy for the arithmetic and wording. Give it your price list and the client's attendance, and ask for a short message showing the per-class cost of the two best-fitting options against the drop-in price. In person, after the last intro class, one sentence is enough: "You've come three times in ten days. If you'd like to keep that rhythm, the twice-a-week membership works out cheapest per class. Want me to send you the details?" One ask in person, one by message, no more.
Check the arithmetic the model does. An illustrative draft for a client who came three times in ten days: "Unlimited at $240 a month works out at just $12 a class!" That figure assumes 20 classes a month, which nobody on an intro offer has shown. Tell the model to base per-class costs on the client's own rate so far (three in ten days is roughly nine a month), and the honest comparison becomes: twice-a-week at $180 is about $20 a class, unlimited at $240 is about $27 at their current pace, drop-in is $32. The recommendation flips to the cheaper plan, and the client can see you worked it out for them rather than for the till.
What your studio software can already automate
Check your platform before building anything. As of September 2026, Momence lists intro-offer conversions among its automated sequences, which it says fire on actions such as a first visit, a lapse or an abandoned checkout, and it has an AI Inbox for clients ready to book or buy. Arketa reports which intro offers convert the most memberships, runs automations, and offers an AI chatbot and phone line. Mindbody says its Messenger[ai] can sell packages and memberships to new clients by text and webchat.
Whatever you use, confirm three things: messages can trigger on attendance, not just purchase date; you can insert an instructor's personal line; and the sequence stops the moment someone buys a membership. The third sounds obvious, and it's the most common embarrassment: a new member receiving "don't miss out on membership" two days after joining.
Test it before real clients see it. Buy the intro offer on a staff test account, attend one class (or have the front desk mark it attended), and check the after-class message arrives. Then buy a membership on that account and wait past the next scheduled step. If the 3-days-before-expiry reminder still turns up, the stop rule is keyed to the wrong event, often "intro offer expired" instead of "any membership purchased". Repeat the test once with the test client answering "yes" to the reformer-experience question, to confirm the experienced journey really does skip the beginner messages.
One reformer studio's funnel, before and after
Say a reformer studio sells 45 intro offers a month. Its three-month funnel averages: 41 attend class 1, 30 attend class 2, 24 use all three classes, and 8 buy a membership within 14 days of expiry. That's 18% of buyers.
- Changes: classes 2 and 3 booked at purchase, the day-5 nudge, a personal after-class message from each instructor, and an attendance-based membership recommendation.
- Suppose class 2 attendance rises to 36 and memberships to 12 a month, about 27% of buyers.
- At $180 a month, and if a member stays about eight months (use your own retention figure), each new member is worth around $1,440. Four extra members a month is about $5,760 from each monthly intake.
- Time cost: about five minutes a day per instructor approving messages, plus a few hours of setup.
These numbers show the sum to run on your own funnel, not a promise. The funnel table tells you whether the change worked at the stage you targeted.
Ways the sequence backfires
- Too many messages. Count everything an intro client receives in two weeks, including booking confirmations. Past about eight, cut.
- A discount at expiry every time. Word spreads and clients wait for it. Use the intro credit instead, and keep it the same for everyone.
- Automation that ignores what the client told you. Someone who mentioned an injury or pregnancy should get a personal message, not the standard nudge; automated check-ins for gym members covers the pause rules that prevent it.
- Invented detail. An AI message praising something the instructor never said is worse than no personal line.
- Chasing people who said they were just trying it. Respect it, and let the post-expiry question be the last contact.
Full classes deserve a look too. An intro client who can't get into the class they liked twice in a row rarely joins; running your class waitlist automatically helps them get a place.
Further reads
- Can AI Answer Gym Enquiries and Book Trial Sessions? — Answer enquiries and book taster classes before the intro offer starts.
- What Is AI Churn Prediction and Can a Small Gym Use It? — Spot new members at risk once they've joined.
- Should AI Handle Membership Freezes and Cancellations? — Handle the freezes and cancellations that follow later.
- How to Win Back Lapsed Customers With AI-Personalised Emails — A later sequence for intro clients who didn't join.
- Gym Retention Software With AI: What It Costs and Delivers — What paid retention tools cost if you outgrow your studio software.
- How to Build a Customer Loyalty Programme With AI Personalisation — Keep new members coming once they've converted.
- How to Clean Up Customer Records Before You Add AI — Four clean-up passes that stop AI emailing people twice, or at all when they said no, with matching rules and a merge log.
- How Cafés Use AI to Fill Quiet Hours With Targeted Offers — Find your café's quiet hours in the till data, design offers for the people free at those times, and measure whether they brought anyone new in.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Momence marketing automation page (sequences on first visit, lapse or abandoned checkout; intro-offer conversions; AI Inbox); Arketa analytics and features pages (intro offer conversion reporting, automations, AI chatbot); Mindbody Messenger[ai] page (selling packages and memberships by text and webchat). Checked September 2026. Studio figures in the example are illustrative.