When we scanned 536 clinic and med spa websites across the Greater Toronto Area, 78.4% had no chat of any kind [1]. Most of the questions that stop someone booking are simple: what does it cost, do you have evenings, where do I park? They often arrive after the front desk has gone home, and they go unanswered.
AI receptionists promise to fix that by answering calls, texts and website chats day and night. Some of that promise is real. Some of it, for a med spa in Canada, needs careful limits. This post covers both.
What an AI receptionist can do well
Answer the same questions your front desk answers every day. Hours, location, parking, price ranges, what to expect at a consultation, how to prepare, cancellation policy. The key is that it answers only from information you have written and approved, not from whatever it finds on the internet.
Catch after-hours enquiries. Someone asking at 10pm about a consultation shouldn't hear nothing until morning. An AI can reply right away, take their name and what they're interested in, and pass it to your team.
Recover missed calls. When a call rings out because everyone is with a client, a text back ("Sorry we missed you. How can we help?") keeps the conversation going instead of losing the caller to the next clinic on Google.
Send people to book. At a minimum, it can send a link to your online booking page. Whether it can make the booking itself depends on your booking software (see below).
Hand off to a person. A good setup knows what it doesn't know. It tells the client a team member will follow up, and it makes sure someone does.
What it shouldn't do
Tell someone whether a treatment is right for them
"Am I a candidate for filler?" "Can I have Botox while breastfeeding?" "Will laser work on my skin type?" These are clinical questions. In Ontario, injectables need an order from an authorized provider, based on an assessment of the client's health history and contraindications [2]. A chatbot isn't that assessment.
The right answer from an AI is: "That's a great question for our [nurse / physician], and it's what the consultation is for. Can I book you one?" That also happens to be the answer that gets the booking.
Describe what a prescription drug will do
Botox and other neurotoxins are prescription drugs. In Canada, advertising them to the public is limited to name, price and quantity [3]. Whatever a chatbot on your website says, it says on your clinic's behalf. So treat its answers like your website copy:
- "Botox is $12 a unit" is fine.
- "Botox will get rid of your forehead lines" is the kind of claim to keep out.
More on the advertising rules.
Collect health details it doesn't need
If someone types "I have melasma and I'm on Accutane, what can I do?", that's health information tied to a person. In Ontario, identifying information about someone's health is personal health information under PHIPA [4]. If your clinic is a health information custodian, those chat logs need the same care as your client records.
The simple rule: keep the AI on admin (names, contact details, times, general questions), and send anything clinical to your team without storing more than it needs.
Pretend to be a person
No Canadian law we're aware of currently requires a chatbot to say it's AI. But clients who find out halfway through a conversation tend to feel misled, and trust matters in aesthetics. Say it up front: "I'm the clinic's virtual assistant. I can answer general questions or get a team member to call you."
Can it actually book appointments?
This depends on your booking software, not on the AI.
| Booking system | Can an AI create the booking itself? |
|---|---|
| Jane | No. Jane's API is for approved partners only, and appointments are read-only [5] |
| Fresha | No. There's no public API for creating bookings |
| Square Appointments | Yes, on the Plus plan or above, through Square's Bookings API [6] |
On Jane or Fresha, an AI receptionist can still help. It answers the question, sends the client your booking link, or takes a request for your team to book. It just can't put the appointment in your calendar itself. If a vendor tells you otherwise for Jane or Fresha, ask exactly how. The answer is sometimes "a person books it for you", which may be fine, but it isn't the same thing.
We compare the three systems in more detail in Med spa booking software: which ones help you rebook clients?.
Following up by text: the consent rules
When an AI follows up with someone who enquired, that's a commercial message under Canada's anti-spam law (CASL). An enquiry gives you implied consent to message that person for six months [7]. After that, you need their express consent. A client who has bought from you gives implied consent for two years from their last purchase [7].
Every message must identify your clinic and include a way to opt out, such as "Reply STOP" [7].
Questions to ask any AI receptionist vendor
- Where do the answers come from? Only from information you approve, or can it make things up?
- What does it do with clinical questions? Ask to see it handle "Am I a candidate?"
- Where are conversations stored, for how long, and who can see them?
- Can it book into my system, or does it send links or requests? (See the table above.)
- How does a hand-off reach my team, and what happens if nobody picks it up?
- Can I read every conversation? You should, at least for the first few weeks.
Where to start
An AI receptionist catches new enquiries. But most clinics already have a bigger, quieter opportunity: past clients who stopped booking.
That's where RelayOps starts every clinic: with a win-back campaign to your existing clients. Missed-call recovery and after-hours answering are added for clinics already running win-back, and scoped with you on a call. The first step is the free lapsed-client report, which shows how many of your clients haven't been back in 3+ months and who to contact first. Get your free report →
Sources
- RelayOps, The State of Med Spa Websites: 536 GTA clinic and med spa websites, scanned 2026-08-14
- College of Nurses of Ontario, Aesthetic Services practice guideline (June 2026), Authority section
- Health Canada, Illegal marketing of prescription drugs; Food and Drug Regulations, s. C.01.044
- Personal Health Information Protection Act, 2004 (Ontario), s. 4
- Jane, Developer Platform API reference (checked 2026-09-30)
- Square Canada, Square Appointments pricing & plans (checked 2026-09-30)
- Canada's Anti-Spam Legislation (S.C. 2010, c. 23), ss. 6 and 10(10)