How AI Engines Choose Which Ontario Dentists to Recommend

Take a new resident in Guelph with a cracked molar and no dentist yet. She does not open Google and scroll a map. She opens ChatGPT and types, "who are the best dentists in Guelph, Ontario." A few seconds later she has a tidy list of five names with addresses and a sentence on each. She picks the second one and books. She never saw a single ranking, never compared two practice homepages, never read a meta description. The recommendation was made for her, upstream, by a machine reading sources she will never look at.

Here is what would surprise the dentist she chose, and the four she did not. The list was not assembled from their websites. We know, because we measured it. Every other guide on this topic is a generic checklist promising that schema markup and a tidy Google Business Profile get a practice recommended. We ran the scrape instead, and it tells a blunter story: the answer routes through directories, and the practice site is mostly a bystander. What follows walks through the signals that actually decide the recommendation, in the order they matter, using real Ontario data rather than borrowed American theory.

A note before we start: this is marketing research, not dental or medical advice. The named practices are real businesses that appeared in our scrape; their inclusion describes how AI engines cited them, not a clinical endorsement or a judgement on care quality.

Why does AI name a directory before it names the dental office?

Because AI engines build a best-dentist answer from third-party directory and listing pages, not from any practice's own site. When a patient asks, the engine runs a retrieval step, pulls a handful of pages it can fetch and trust in that moment, reads the names inside them, and writes the list. A directory that ranks ten dentists with addresses and a one-line strength each is far easier to ground than a polished practice homepage written as marketing prose. So the directory wins the citation, and the practice site sits unread.

The academic literature has already given this behaviour a formal label. Pranjal Aggarwal and colleagues described it in "GEO: Generative Engine Optimization" (arXiv:2311.09735), presented at KDD 2024 by ACM SIGKDD. Generative engines answer by synthesizing and summarising several cited sources rather than returning a ranked list of links, and the paper showed that adding citations, quotations, and statistics to a source can raise its visibility inside AI answers by up to 40 percent. Being a citable source is the lever. Holding Google's first organic position for a dental keyword is not. A practice can hold position one for "dentist Guelph" and still lose the AI answer to a directory page that the engine found easier to read.

Kevin Indig's Growth Memo from March 2026 adds the positional rule underneath it: about 44 percent of AI citations come from the first 30 percent of a page. A directory leads with a clean ranked list in its first screen. A practice "Meet the Team" page buries its services under staff biographies, giving the engine nothing extractable near the top. The directory is simply built the way the engine likes to read.

Each engine grounds its dentist answer in a different slice of the web

The four engines do not read the same web, so "getting recommended by AI" is really four separate jobs. We asked ChatGPT (OpenAI), Anthropic Claude, Google Gemini, and Perplexity the same question for nine Ontario cities, Burlington, Cambridge, Guelph, Hamilton, Kitchener, Mississauga, Oakville, Toronto, and Waterloo, and logged every cited source. Across the full Formative Digital dataset of 583 distinct domains over 44 city and vertical cells, only 95 sources (16.3 percent) were cited by two or more engines, so more than 83 percent were unique to a single engine. For dentists, that divergence shows up as a clear per-engine personality.

Read those four behaviours side by side and the takeaway writes itself. If a Hamilton practice optimised perfectly for the sources ChatGPT trusts, it would have done almost nothing for Claude, which never looked at Google and went straight to ThreeBestRated and Opencare. Four engines mean four reading lists, and we unpack that pattern across every vertical in our study of the 16.3% cross-engine overlap problem.

Which directories did AI cite most for Ontario dentists?

A short roster of platforms supplied most dental citations across all nine cities, and those names are the practice owner's actual battleground. If your clinic is absent from these, it is absent from the layer the engines read. If it is present and well-described, it has a real path into the answer.

One detail rewards a second look. Opencare and Google Maps are the only non-Claude sources to recur across many cities. Opencare has the broadest cross-engine reach for dental, while a complete Google Business Profile is what feeds ChatGPT specifically. The Claude-favoured listicles (ThreeBestRated, goodcaring, RateMDs) are high-volume but single-engine. So a practice chasing all four engines prioritises differently than one chasing only the engine its patients use. The concentration is not unique to dentistry: a 2025 study reported by Search Engine Land found that within a topic, roughly 30 domains capture about two-thirds of ChatGPT's citations (Search Engine Land). The seats are few, and in Ontario dental the directories hold most of them. We dig into the Maps mechanism on its own in our look at the local signals that decide AI near-me answers.

Do reviews, schema, and a Google Business Profile actually move the answer?

They help, but only on the engines and in the layers where they are actually read, which is the nuance the generic guides oversell. They are not a master switch that makes every engine name you. They are inputs into specific source layers, and their value depends entirely on which engine you are trying to reach.

Each of these makes your practice easier to read and attribute inside the layers a particular engine grounds against, and that is worth the effort. It is not the same as the promise, repeated across most dental-marketing blogs, that ticking the schema and Google Business Profile boxes makes the engines name you. Our data shows the recommendation flowing through directories the practice does not control, which is why the work cannot stop at the practice's own site.

The RCDSO problem hiding inside a "best dentist" citation

A compliance tension sits at the centre of this vertical: the directories an AI engine cites call Ontario practices the "best dentist," a superlative the Royal College of Dental Surgeons of Ontario forbids the dentist from claiming. This is where dental GEO stops being a pure marketing question and becomes a Your Money or Your Life one, because patients act on these recommendations when choosing a health provider.

The stakes are not abstract, because effectively the entire patient population is now reachable this way. DataReportal's Digital 2025: Canada report counted 38.0 million internet users at a 95.2 percent penetration rate. A growing share of those people ask an engine, not a friend, for a dentist. When the recommendation arrives wrapped in a word the regulator forbids the provider to use, accuracy and restraint are the compliant path, and they happen to be the citable one too.

What can a single-location Ontario practice actually do about it?

A single-location practice wins AI visibility by earning a clean, accurate presence in the directories the engines already pull, then structuring its own site so the model can read and attribute it. You are not trying to outrank a competitor on Google. You are trying to be present and legible in the specific sources each engine grounds against. That is a narrower, more achievable job than chasing a national algorithm, and it maps directly onto the directories our scrape surfaced.

This is where two of Formative Digital's 12 Vectors carry the load. Vector 5, Cite, earns placement in the third-party sources each engine trusts, which for dental means Opencare, ThreeBestRated, RateMDs, and the rest. Vector 10, Localize, binds your practice name, address, and clinic identity into one consistent record so no engine mistakes you for another Ontario dental office. We run these through the Formative Forces, our orchestrated multi-agent system, so one practice is worked across all four source layers in parallel rather than one engine at a time. The same mechanics in a different field sit in our retail what FD's content engine did for a Brantford retailer, where structured, citable content moved a Brantford business from roughly 1,000 to more tha 91,700 monthly organic visits (SEMrush, April 2026). Dental is a harder, double-YMYL vertical, and outcomes depend on your competition, your existing presence, and your city, so treat that as direction, not a promise.

What should the practice's own site carry?

The directory layer decides whether an engine finds you. Your own site decides what it can say once it has, and what a patient sees when they click through. Most practice sites have one Services page listing every treatment as a bullet point, which gives an engine one vague page to cite for twenty different questions. The build we recommend is narrower and deeper.

On the markup side, Dentist is the specific Schema.org type for the clinic (it sits under both LocalBusiness and MedicalOrganization), its isAcceptingNewPatients property tells a machine whether you are booking, and each treatment page can carry a MedicalProcedure node as the page's subject. Connect them with @id references so the clinic, its dentists and its treatments resolve as one practice.

Where RCDSO rules shape the build

RCDSO's advertising guidelines reach further into page design than most agencies realise (RCDSO Advertising Guidelines). Four rules change what a treatment page can contain:

• Any advertisement that refers to a dental area, procedure or treatment must clearly disclose whether you are registered as a general practitioner or a specialist, and if a specialist, in which specialty. That belongs on every treatment page, near the top.
• Testimonials are prohibited, along with any statement that can only be verified by someone's personal feelings. Do not paste Google reviews onto your treatment pages; earn them on Google, Opencare and RateMDs, where the engines read them.
• Incentive programs are out: giveaways, contests, draws, free products or services. That rules out the "free whitening with your first exam" offer, and it is a good reason never to reward patients for leaving a review.
• If you publish a fee, it must be clear, and unless you say otherwise it is read as the maximum fee, inclusive of all services including lab costs, for every patient regardless of insurance.

The guidelines also prohibit statements that suggest or create an expectation of favourable results, which is a useful test for any line of treatment copy before it ships.

How do you measure dental AI visibility across four engines?

You track it per engine, not as a single score, and most practices get it wrong by chasing the wrong number. You end up with four numbers, one per engine, and they will disagree. The method is to run the real patient query, "best dentists in {your city}, Ontario," through ChatGPT, Claude, Gemini, and Perplexity on a schedule, recording which practices each engine names and in what order, and tracking your share of those mentions over time. A Google ranking report tells you almost nothing about this, because the engines barely read Google's ranked results.

Two refinements matter. First, watch the source layer as well as the answer, because the layer is the leading indicator. If you newly appear on ThreeBestRated, expect movement in Claude before anywhere else; the listing changes first and the citation follows. Second, sample each engine more than once, because AI answers carry run-to-run variance, so a single check on a single day is weak even for one engine. Track four engines as four scores, sampled repeatedly, and the picture becomes honest. We lay out the full method in our guide to diagnosing your AI visibility across engines.

None of this is magic ranking dust, and nobody can promise a given engine will name a given practice on a given day. The engines shift, the directories reshuffle, and a health vertical is among the most cautious. What the data supports is direction: the citation goes to retrievable, attributable, compliant sources, so a practice that earns its place in the directories the engines already trust, and keeps its own house clean and readable, competes for the AI answer far better than one still polishing a Google ranking the engines never open.

Who this work fits, how long it takes, and what it costs

Formative Digital runs this as a service for Ontario dental practices. It fits single-location and small-group practices with an existing domain, genuine patient trust that does not yet show up in AI answers, and a clear focus (family, paediatric, implants, sedation, emergency care) that a flat directory listing erases. It does not fit anyone who wants a guaranteed number-one spot or a promise that every engine will name the practice, and we will not write superlative or comparative copy, or coach an engine to repeat it, because RCDSO prohibits it.

Timelines run on two clocks. Google AI Overviews and Perplexity read the live web, so corrections can show there within 30 to 60 days; ChatGPT and Gemini lean more on slower update cycles and typically take three to nine months. In the local pack, plan on 30 to 45 days for early movement in smaller markets and 90 to 120 days in competitive metros, and twelve to eighteen months for the "dentist plus city" head term on a domain younger than three years. Those ranges vary with your city and your starting point.

Our published tiers are Starter, Growth and Dominance, listed with current figures at /pricing/; most single-location and small-group practices fit Starter or Growth, and groups competing across a region tend toward Dominance. Everything is month to month with a written cancellation clause. Results Guarantee: if your existing domain shows no measurable organic search results after 12 months of work with Formative Digital, we work for free until you see them.