Should You Use AI in Your Ads? Green Flag, With One Hard Exception
- Andrew Vass
- 10 minutes ago
- 5 min read
The assumption doing the rounds is that AI in advertising is either fine everywhere or a scandal everywhere.
It's neither, and treating it as a single question is why clinics keep getting this wrong.
I run a marketing agency. I use AI every day. I'd also tell any wellness clinic owner that one specific use of it will cost them bookings, and it happens to be the use most people reach for first.
Watch the short version here: Should You Use AI in Your Ads?. You can also catch it on Instagram.
The answer depends entirely on your industry
If you're selling software or an e-commerce product, most buyers genuinely don't care that an ad was generated.
Someone deciding on a project management tool or a pair of headphones isn't buying a relationship. If they clocked that the image was AI, or could just tell, it wouldn't move their decision much. The purchase doesn't depend on trusting a specific human.
Wellness is a different transaction.
People are handing you their health. They're deciding whether to be honest with you about symptoms they've been embarrassed about for two years. That decision runs entirely on trust in a person, and an AI-generated face doesn't build trust. It usually does the opposite.
That's the whole distinction. Not "is AI acceptable" but "does this purchase depend on believing a specific human is real."
Why a generated face actively backfires
It's worth being precise about the mechanism, because "it feels off" is the entire effect.
Near-human faces that aren't quite right produce discomfort rather than neutrality. Researchers have called this the uncanny valley since the seventies: the closer a synthetic figure gets to human without arriving, the more unsettling it becomes.
So an AI presenter doesn't land as "a slightly less good ad." It lands as something subtly wrong, and the viewer doesn't always know why.
Then a second effect stacks on top. Once someone notices the person isn't real, every claim behind that person gets discounted. If the face was fake, what else was.
We've watched this play out in a real clinic campaign. One we reviewed opened with an AI-generated presenter saying "Hi, I'm Dr. [name]." We rated it 4 to 5 out of 10, and the presenter was the single biggest problem. The clinic was legitimate. The ad introduced it with a person who didn't exist.
Compare that to the highest-rated ad we've reviewed, an 8 out of 10, which was a real practitioner explaining one problem to camera with no production at all.
Where AI earns its place
None of that is an argument against using AI. I use it constantly, and there's a clear line between the uses that help and the one that hurts.
Use AI for | Don't use AI for |
Writing and refining scripts | The person on camera |
Generating angles and topics | Fake testimonials or patient stories |
Drafting variations to test | Invented results or numbers |
Research and summarising | A voice pretending to be the practitioner |
Editing, captions, repurposing | Anything implying a person exists who doesn't |
Everything in the left column is invisible to the patient and makes the work better. Everything in the right column is a claim about who you are.
The dividing line isn't how advanced the tool is. It's whether the output misrepresents a human relationship the patient is being asked to enter.
The uncomfortable part
The reason clinic owners reach for AI avatars isn't laziness. It's that being on camera is uncomfortable, and a generated presenter is a way to have video without having to do that.
I understand the impulse and I'd still say don't.
If you're a naturopathic doctor or a wellness coach, you need to learn to be good on camera and do talking videos. It's the most personal thing you can do, and it's the only thing that builds trust and credibility at the speed this business needs.
Your first few will be bad. Everyone's are. That's the cost of getting to the ones that aren't, and there's no tool that lets you skip it, because the thing being purchased is you.
We've written more about that trade in why looking perfect costs you bookings.
What to do this week
Audit your current creative for synthetic humans. Avatars, generated presenters, stock models posing as patients. Replace them with you.
Keep the AI where it's invisible. Scripts, hooks, topic lists, editing. That's leverage with no trust cost.
Film one ninety-second video on your phone. No script, no set. Explain one condition the way you'd explain it in the room.
Never generate a testimonial. Not a paraphrase, not a composite, not an illustrative example. Real patients or nothing.
Ask one patient to watch your ad and tell you if anything feels off. They'll spot it faster than you will.
Frequently asked questions
Should wellness clinics use AI in their ads?
Yes for the invisible work, no for the person on camera. AI is genuinely useful for scripts, ideas, topics and editing. An AI-generated presenter or avatar undermines the trust the booking depends on, because patients are choosing a practitioner rather than a product.
Why do AI avatars hurt clinic ads specifically?
Because the purchase depends on trusting a specific human with your health. Near-human faces that aren't quite right create discomfort rather than neutrality, and once a viewer notices the person isn't real, every claim behind them gets discounted.
Is it different for e-commerce or software companies?
Yes. If someone is buying a product rather than a relationship, most buyers don't care whether the creative was generated, and it rarely changes the purchase decision. The trust requirement is what makes wellness different, not the technology.
Can I use AI to write my ad scripts?
Absolutely, and it's one of the best uses of it. Scripts, hooks, angles and topic lists are all invisible to the patient and the tool makes that work faster. Keep your own voice in the final pass so it still sounds like you in the room.
Is it ever acceptable to generate a testimonial?
No. That's a claim about a real person's experience that didn't happen, and it's the fastest way to lose both trust and, potentially, regulatory standing. Real patients, real words, with permission.
Does Lead Shark Marketing use AI when building clinic campaigns?
Yes, every day, for scripts, ideas and topics. What we don't do is put a generated face in front of a wellness audience, because the thing being sold is the practitioner.
Want to see exactly how wellness clinics are filling their schedules on autopilot? Andrew runs a free training breaking down the system Lead Shark uses to book new patient appointments every month for clinics like yours, without posting on social media every day or chasing leads by hand. Register for the free training
The short version
Whether AI belongs in your ads depends on whether the purchase depends on trusting a person.
Selling software or a product, nobody much cares. Selling care, an AI face does the opposite of what you needed it to do.
Use AI for scripts, ideas, topics and editing. Put your own face on camera.
That's the whole rule, and the discomfort of being on camera is the price of the trust that gets people booked.
Ready to stop guessing at marketing and start getting predictable new patients?
Lead Shark Marketing builds done-for-you patient acquisition systems for wellness clinics, naturopaths and health practitioners across North America. If you want to know whether it's a fit for your practice, start here:
Andrew Vass runs Lead Shark Marketing, where he builds lead follow-up and appointment-setting systems for wellness and medical clinics across Canada. He spends most of his time watching clinics lose booked revenue in the gap between a lead coming in and someone calling them back.
