ChatGPT Ads and Canadian Healthcare

Somewhere in your inbox there’s probably a pitch about advertising on ChatGPT (we sent one out to our clients not too long ago as well 😅). Ads showed up in ChatGPT in the US in February, they reached Canada in the spring, and by the end of August OpenAI was reporting a billion dollars in annualized ad revenue.[1] So the pitches were always going to come.

I run paid media at Consultus Digital, including our healthcare accounts, and I’ve been getting the question from clinic owners for a few months now. The short answer is that a Canadian clinic can’t buy ads on ChatGPT today. OpenAI’s own policy says health services ads outside the US are “generally prohibited,” and the exceptions it has started making are for US advertisers only.[2]

The longer answer is worth your time, because your patients are already in there asking about their symptoms, and the rules that keep you off the platform today also tell you what to prepare for when that changes.

I’ve written this for clinic owners rather than marketers, so I’ve kept the platform mechanics to what you’d need to make a decision. Everything quoted below comes from OpenAI’s own pages, Google’s, or a named survey, and the footnotes have the dates. OpenAI has rewritten its ad policy several times this year, so check the policy page again on the day you read this. Or shameless plug, you can reach out to us, this is what we do. 

What a ChatGPT ad actually is

If you pay for ChatGPT, you’ve never seen one. OpenAI’s help centre says ads “may appear for users on the Free and Go plans,” and that “Plus, Pro, Business, Enterprise, and Edu accounts will not have ads.”[3] They sit below the end of a response, labelled as sponsored and visually separated from the answer itself.[3]

That placement is the interesting part for a clinic. A search ad shows up next to a list of links. A ChatGPT ad shows up under an answer the person just read, which means the ad is only as relevant as the conversation it followed.

Two other rules matter more than any targeting option. OpenAI doesn’t show ads to accounts it identifies as belonging to users under 18, using account-level age information and age prediction where available.[3] And ads “are not eligible to appear near sensitive or regulated topics, including personal health, mental health, or politics.”[3]

Read that last one twice. Even if a clinic were an approved advertiser, its ad wouldn’t appear under the conversation where someone describes their back pain. The health conversation is one of very few placements OpenAI has fenced off. The ad policy goes further and lists “Mental and personal health conversations” among the vulnerable interactions that get no ads at all.[2]

So the clinic that imagines its ad appearing right after ChatGPT says “you should see a physiotherapist” is imagining something OpenAI has said it won’t do. I’d want that understood before anyone spends a dollar.

Can a Canadian clinic advertise on ChatGPT today?

The answer is no, with one narrow footnote I’ll get to.

OpenAI’s ad policy, last updated September 10, 2026, puts it this way: “Ads for regulated medical products and services, or claims related to the prevention, diagnosis, or treatment of physical or mental health conditions, are restricted. Ads for health services outside the US are generally prohibited.”[2]

Inside the US, the door has opened a crack. The same policy says OpenAI “may allow ads from approved advertisers on a case-by-case basis” in nine categories:[2]

• Consumer medical devices, wearables and apps
• Dental services
• Dietary supplements
• Disease awareness campaigns that don’t advocate for public policy
• Health insurance
• Hospitals and urgent care
• Medical testing and diagnostic products and services
• Minimally invasive cosmetic procedures
• Vision products and services

If you run a dental office in Buffalo, you can apply. If you run one in Mississauga, you can’t, and the only difference is which side of the border you’re billed from. The policy also notes that “advertisers may be required to provide proof of licensure,” so the US path isn’t a self-serve checkbox either.[2]

A few things aren’t on that list at all: prescription drugs, pharmacies, therapy and counselling, telehealth and veterinary clinics. Marketing Brew reported in June that pharmaceutical brands weren’t permitted in the pilot and that Havas was pushing OpenAI to change that.[4] For the rest, the general restriction on “regulated medical products and services” applies.

The narrow footnote

The word in the policy is “generally.” OpenAI reviews health advertisers manually, and “generally prohibited” leaves room for a decision I can’t predict. If you want a definitive answer for your clinic, the only way to get one is to apply and be told. I wouldn’t budget for a yes.

Where Canada stands on the platform itself

The health rule is the wall. Everything else about Canada is open, which is what makes the wall frustrating.

OpenAI announced on March 26 that it would begin “expanding beyond the U.S., starting with pilots in Canada, Australia, and New Zealand.”[5] Trade coverage put the Canadian go-live in mid-April, though OpenAI’s own pages don’t give a day.[6] By May, Digiday was listing Canada among the live markets while OpenAI added the UK, Brazil and Japan.[7] We’re live with a few non-health clients today, all seeing some promising initial results.

Canada is now on the list of countries where a business can open a self-serve Ads Manager account. The one condition OpenAI states is that “the legal entity that will advertise and be billed must be based in a country listed as available.”[8] A Canadian corporation qualifies.

The mechanics, for anyone comparing this to Google or Meta:

Setting	What OpenAI’s help centre says	Source
Minimum daily budget	25 CAD	[9]
Buying options	Views (CPM), Clicks (CPC), or Conversions	[9]
Geo targeting	Country, plus “states or regions, cities, markets, and postal codes where available”	[9]
Custom audiences	Inclusion audiences need “at least 25,000 matched users”; smaller lists can only be used for exclusions	[10]
Who sees ads	Free and Go plan users, 18 and over	[3]

Two of those rows would matter to a clinic even if the health rule vanished tomorrow.

The audience minimum rules out any patient-list targeting for a normal practice. A clinic with 4,000 patients on file is nowhere near 25,000 matched users, and matching a patient list to an ad platform raises the same privacy questions I covered in my last article, so I’d treat that row as closed on two counts.

Your patients are in there anyway

None of the above would matter much if ChatGPT were a niche tool. It isn’t, and the health numbers are the ones that should get a clinic owner’s attention.

OpenAI’s own report from January says “more than 40M turn to ChatGPT every day with healthcare questions,” that “more than 5% of all ChatGPT messages globally are about healthcare,” and that “seven in 10 healthcare conversations in ChatGPT happen outside of normal clinic hours.”[11] Those are OpenAI’s figures about OpenAI’s product, so read them as a company describing itself. Even discounted, they’re large.

The Canadian data is more sobering. In February, the Canadian Medical Association released survey results from Abacus Data showing that 89% of Canadians go online for health information, that “about half are using AI tools to diagnose or treat their health issues,” and that “only a quarter of Canadians (27%) trust AI to provide accurate health information.”[12] The CMA’s release doesn’t give a sample size, so I’d cite the direction rather than the decimal.

The same release reports that Canadians who followed health advice from AI were “five times more likely to experience harms than those who did not.”[12] A peer-reviewed study in Nature Medicine the same week, with nearly 1,300 participants, found that people using AI chatbots “did not make better decisions than participants who relied on traditional methods like online searches or their own judgment.”[13]

For a US comparison, KFF’s March poll of 1,343 adults found 32% had turned to AI for health information and advice, 16% for mental health specifically, and 77% were concerned about the privacy of medical information they share with AI tools.[14]

Put those together and you get a picture of a patient who asks ChatGPT first, doesn’t fully trust the answer, and then goes looking for a human. That last step is where a clinic still has a role, and it happens on Google, on your website, and on the phone. It doesn’t happen inside ChatGPT, because OpenAI has decided that health conversations get no ads.

Showing up inside the answer without buying an ad

The ad slot under a health conversation is closed. The answer itself is a different story. When someone asks ChatGPT for a physiotherapist in Etobicoke or a dentist that takes new patients in Barrie, ChatGPT searches the web and names clinics, and no clinic paid to be there. Getting into that list is the work I’d spend time on this year.

The industry has taken to calling this generative engine optimization, or GEO. Strip the acronym away and it’s the same discipline as search engine optimization, pointed at a different reader. At Consultus we treat it as an extension of our SEO work rather than a separate product, and it’s something we help clinics with. Feel free to book some time with use to explore how we can help you still appear within not only ChatGPT but all AI platforms.

What OpenAI has actually published

OpenAI says very little about how ChatGPT picks sources, so I’ll start with what it does say.

Its publisher FAQ states that “any public website can appear in ChatGPT search,” and that the main thing a site owner controls is whether OpenAI’s search crawler, OAI-SearchBot, is allowed in.[21] OpenAI’s crawler documentation says OAI-SearchBot “is used to surface websites in search results in ChatGPT’s search features,” that site owners allow or block it in their robots.txt file, and that changes take about 24 hours to register.[22] A site that blocks it won’t be used in search answers.

The same FAQ gives you a way to measure this. ChatGPT “automatically includes the UTM parameter utm_source=chatgpt.com in referral URLs,” so visits from ChatGPT show up in your analytics under that source.[21] Before anyone sells you a GEO package, open GA4, filter on that source, and see whether you’re getting any of this traffic already.

OpenAI does not publish a ranking method, a list of signals, or a way to pay for placement in search results. Anyone who tells you they know the formula is guessing, including me.

What we do, and why

Because there’s no published formula, the work is inference from how these systems behave, and I’ll label it as such. When ChatGPT answers a local health question, it reads web pages and summarizes them. A page it can read, that states the facts a patient asks about, gives it more to work with than a page built around a hero image and a slogan. That’s the whole theory, and it’s the same theory that made SEO work for twenty years.

In practice that means:

• Let the crawler in. Check robots.txt for OAI-SearchBot and confirm nobody blocked it as part of an AI-training opt-out. Blocking GPTBot, the training crawler, is a separate decision and doesn’t affect search.[22]
• State the facts in text. Services, conditions treated, locations, hours, whether you’re accepting new patients, direct billing, languages spoken and what a first visit costs. In sentences on the page, not in a PDF, a slider or an image.
• Answer the questions patients actually ask. The CMA survey above says half of Canadians are using AI to diagnose or treat. Those are symptom questions, so a service page that plainly explains what a condition is and when to book has a job to do here.
• Keep your Google Business Profile and directory listings consistent. An AI that finds three different phone numbers for you has a reason to name someone else.
• Track it. Set up the utm_source=chatgpt.com view in GA4 and report it monthly next to organic search.

Everything on that list is inferred from how the systems work, not promised by OpenAI. What I can say with confidence is that none of it hurts your Google rankings, so the downside is staff time.

The rules that follow you onto any platform

When OpenAI eventually opens health ads to Canadian advertisers, and I’d expect it to at some point, the platform’s policy will be the easier half of the compliance work. The Canadian rules don’t care which platform you’re on, and a conversational ad format makes some of them harder to follow, not easier.

I’m not a lawyer, so treat this as a list of things to check with whoever handles compliance at your clinic.

Health Canada

The Food and Drug Regulations say that anyone advertising a prescription drug to the general public “shall not make any representation other than with respect to the brand name, the proper name, the common name and the price and quantity of the drug.”[15] The Food and Drugs Act adds that no one may advertise a drug or device to the public “as a treatment, preventative or cure” for the conditions listed in Schedule A.[15]

A chat interface invites a back-and-forth. A pharmacy or a clinic offering a named medication would need its ad to say almost nothing, and OpenAI’s US allowances don’t even cover pharmacies yet.

Your regulatory college

Every regulated profession in Ontario has advertising rules, and they’re stricter than most owners remember. Two examples:

The College of Physicians and Surgeons of Ontario says physicians must not advertise in a manner that “contains a testimonial,” or that “promises or suggests a better or more effective service than any other physician or health professional.” It also says physicians “must ensure that any advertisement prepared by them, or on their behalf by a third party, complies” with the policy.[16] That last line means your agency’s mistake is your problem.

The Royal College of Dental Surgeons of Ontario bans “superlative or comparative terms, such as ‘state of the art,’ ‘cutting edge’ or any other words or phrases to suggest a higher quality,” as well as testimonials and “incentive programs, including giveaways, contests, draws or free products or services.”[17]

Those rules already shape what we can write in a Google ad. They’d shape a ChatGPT ad the same way, and an ad sitting under a friendly conversational answer would probably tempt a copywriter toward exactly the kind of language the colleges prohibit.

What a clinic can actually buy in AI search right now

ChatGPT isn’t the only place people ask an AI about their health, so it’s fair to ask what the other platforms allow.

Google AI Overviews

Google’s ads in AI Overviews are live in Canada. Its help page lists availability “in English on mobile and desktop devices” in Australia, Canada, India, Indonesia, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, Philippines, Singapore and the US.[18] The same page says: “Currently, we don’t show ads in AI Overviews for sensitive verticals like adult, alcohol, gambling, finance, healthcare, politics, and more.”[18]

So Google has drawn the same line OpenAI has. A Canadian clinic’s search ads still run in the regular results below the AI answer, and that’s where they’ve been running all along. 

Microsoft Copilot

Microsoft takes a different approach. Its documentation says “if an ad can serve on Microsoft Bing, it can also serve in Copilot,” that eligible campaigns are opted in automatically, and that advertisers “cannot opt out of serving ads.”[20] 

The part you control

That leaves the unpaid side, which is the section above on showing up inside the answer. It’s the part a clinic can act on this month, and it costs staff time rather than ad spend.

What I’d do this quarter

None of this requires a ChatGPT Ads account, which is convenient, since you can’t use one.

• ☐ When an agency pitches you ChatGPT ads, ask them to show you the line in OpenAI’s ad policy that permits a Canadian health advertiser. If they can’t, you’ve learned something about the agency.
• ☐ Ask your front desk to keep a tally for a month of new patients who mention they asked an AI first. It’s a small number to collect and it tells you whether this is theoretical for your practice.
• ☐ Read your own website the way a chatbot would. Does it state your services, location, hours, new-patient status and typical first-visit cost in plain words on the page? If any of that lives only in a PDF or an image, fix it.
• ☐ Check your Google Business Profile for the same facts. It’s the most likely thing an AI answer draws on when someone asks for a clinic near them, and it’s free to correct.
• ☐ Pull your college’s advertising standard and keep it next to whoever writes your ads. Any ChatGPT ad you eventually run will have to pass the same rules.
• ☐ If you run Microsoft Ads, ask whoever manages them whether your campaigns are showing in Copilot and what those impressions look like.
• ☐ Put a note in your calendar to re-read OpenAI’s ad policy page in three months. It has changed several times this year, and the section that matters to you is one paragraph long.
• ☐ Open yoursite.ca/robots.txt and look for OAI-SearchBot. If it’s disallowed, find out who did that and why before you change it.
• ☐ Filter GA4 for the source chatgpt.com and note the number, even if it’s zero. That’s your baseline.

There’s one caveat before you spend time on any of this. If your clinic is full and you’re turning patients away, the AI question can wait. This is a planning exercise for practices that need growth, and there’s no penalty for being second onto a platform that doesn’t allow you on yet.

Where this leaves a Canadian clinic

ChatGPT ads are real, they’re in Canada, but a healthcare clinic can’t buy them. Patients are asking ChatGPT about their health in large numbers, and OpenAI has decided those conversations get no ads from anyone. Both of those things could change, and the US case-by-case list suggests the direction. I’d still expect Canadian clinics to wait a while, and to face their own college’s rules on the far side of it.

Until then, the patient who asks an AI first still ends up on Google or on your phone line. That’s the part of the journey a clinic can measure and pay for today, and it’s where I’d keep the budget.

If you’d like a second opinion on where your clinic’s ad budget should sit while all this shakes out, the Consultus Digital Healthcare Division offers a free strategy call.

Footnotes

Undated help pages are cited with the date they were accessed. OpenAI’s help centre shows only relative dates such as “updated 3 days ago,” so the access date is the only fixed one.

1. OpenAI. “A milestone in expanding access to AI.” August 31, 2026.
2. OpenAI. “Ad policies.” Updated September 10, 2026. Accessed September 25, 2026.
3. OpenAI Help Center. “Ads in ChatGPT.” Accessed September 25, 2026.
4. Sheena J. “ChatGPT is making room for ads from regulated verticals, but most marketers aren’t crossing the threshold yet.” Marketing Brew, June 17, 2026.
5. OpenAI. “Testing ads in ChatGPT.” February 9, 2026, with dated updates through August 2026.
6. Salvi A. “ChatGPT Ads Go Live in Canada, Australia, and New Zealand.” TechWyse, April 20, 2026. TechWyse is a Toronto agency, so treat the go-live date as press-reported rather than confirmed by OpenAI.
7. Scanlon K. “‘Expand thoughtfully’: OpenAI offers ChatGPT ads to new markets, including the U.K., Brazil and Japan.” Digiday, May 7, 2026.
8. OpenAI Help Center. “Ads Manager Availability.” Accessed September 25, 2026.
9. OpenAI Help Center. “Create Campaigns for ChatGPT Ads.” Accessed September 25, 2026.
10. OpenAI Help Center. “Set up Custom Audiences for your Campaign.” Accessed September 25, 2026.
11. OpenAI. “AI as a Healthcare Ally.” January 2026.
12. Canadian Medical Association. “Doctors warn: Canadians are turning to AI for health information and it is hurting them.” February 10, 2026. Survey by Abacus Data; the release does not state a sample size.
13. University of Oxford. “New study warns of risks in AI chatbots giving medical advice.” February 10, 2026, reporting a study published in Nature Medicine.
14. KFF. “KFF Tracking Poll on Health Information and Trust: Use of AI for Health Information and Advice.” March 25, 2026. Fieldwork February 24 to March 2, 2026; 1,343 US adults.
15. Health Canada. “Relevant legislative and regulatory sections: Guide on distinction between advertising and other activities for health products.” Date modified July 31, 2023.
16. College of Physicians and Surgeons of Ontario. “Advertising.” Approved December 2020.
17. Royal College of Dental Surgeons of Ontario. “Advertising Guidelines.” Accessed September 25, 2026.
18. Google Ads Help. “About ads and AI Overviews.” Accessed September 25, 2026.
19. Google. “A new generation of ads for the AI era of Search.” May 20, 2026.
20. Microsoft Advertising. “About ads in Copilot.” Last updated June 19, 2026.
21. OpenAI Help Center. “Publishers and Developers – FAQ.” Accessed September 25, 2026.
22. OpenAI. “Overview of OpenAI Crawlers.” Accessed September 25, 2026.