82%
of Google's top-10 practices were named by at least one AI engine
We pulled the Google Maps roster for "dentist" in five U.S. cities, then asked three AI assistants who they'd name. The two lists overlap far less than you'd hope — and the gap is teachable.
For twenty years, "getting found" meant one thing: rank on Google. A patient typed "dentist near me," scanned the map pack, and called someone near the top. Local SEO — reviews, proximity, a tidy Google Business Profile — was the whole game.
That game still exists. But a second one has opened next to it. A growing share of people now ask an assistant — ChatGPT, Gemini, Perplexity — "who's a good dentist in my town?" and act on the handful of names it returns. The assistant doesn't show a map of ten options. It hands over a shortlist, often three to five names, in confident prose.
Ranking well on Google no longer guarantees you exist in the answer an AI gives. Winning that answer is AEO — and it follows different rules than SEO.
This study set out to measure the gap: where the AI shortlist agrees with Google's roster, where it diverges, and — most usefully — what seems to move a practice from invisible to named.
We ran the study in five U.S. cities — McKinney TX, Arlington Heights IL, Manhattan Beach CA, Hoboken NJ, and Naperville IL. They aren't the biggest names on the map, and that's deliberate: they're affluent, high-competition local dental markets, which is exactly where ranking is hardest and most worth measuring. They also mirror how people actually search — by their own city or neighborhood ("dentist in Naperville," "dentist in Santa Monica"), not by a giant metro ("dentist in Chicago") — so each one is a bounded local market with a real field of competitors.
For each city we ran a single Google Places query for "dentist [city]" with location bias off, and recorded the practices in three layers: Stratum 1 (the top ten), Stratum 2 (roughly Maps positions 11–20), and Stratum 3 (supplementary results further down, plus any practice an AI named that Google's top tiers had missed). That's 118 practices in total.
Separately, we asked three assistants — ChatGPT, Google Gemini, and Perplexity — to name dentists in each city, and recorded who they listed and in what order. (We also captured a Claude column but excluded it from the analysis, since it isn't a blind comparison.)
It's a snapshot. AI answers drift — run to run, day to day, and in steps when models update. Treat any single practice's presence or absence as indicative, not permanent. The takeaways here are the patterns that held across all five cities, not any one name.
It's Places order, not the local pack. The Google position we recorded is the Places API's prominence order for the bare query — not the proximity-personalized map pack a specific searcher sees from their kitchen. It's a clean, repeatable baseline, not a claim about any one neighborhood.
The single clearest pattern: AI visibility doesn't fade gradually as you go down Google's list. It drops off a ledge after the top ten.
A practice in Google's top ten was named by AI almost twice as often as one sitting just below it. The assistants behave as if there's a credibility threshold around the first page — cross it and you're a candidate; fall just short and you mostly vanish from the conversation, no matter how good you are.
You can see the same threshold in where AI ranks the names it does mention. Among practices the engines listed, the ones from Google's top ten landed around position 4.3 in the AI's own list; those from the next tier landed near 6.6 — roughly a two-slot penalty for starting lower.
If top-ten placement is the gate, the obvious question is what's behind the gate. We tested the usual suspects. The results are blunt.
Across the practices with a real Maps rank, the ones AI named sat at an average position of 8.5; the ones it skipped averaged 13.2. Nothing else we measured separated named from unnamed as cleanly.
Read carefully, Maps position is best understood as a composite proxy: Google has already folded reviews, profile completeness, relevance and prominence into that ordering, and the AI engines — especially the ones that search live — lean on it.
The most consistent through-line among the practices that ranked first in an AI list was having the town or city name in the practice name ("Naperville Family Dentists," "Arlington Heights Dental Care"). Raw review counts, by contrast, were close to flat — and sometimes inverted — within a tier. Two of the biggest review leaders in our set, both without a city name, underperformed badly in the AI lists. Reviews appear to help you clear Google's gate; once you're past it, they stop predicting much.
Two findings genuinely surprised us. First, third-party "authority" — being written up in roundups, press, forums — did not explain which practices AI surfaced. Almost every practice in the study scored at the floor of our external-authority scale (directory listings only; not one genuine editorial roundup appeared for the entire supplementary tier). Second, owning a keyword-perfect domain didn't help: one Naperville practice owns dentistnapervilleil.com — about as on-the-nose as a dental domain gets — and no engine named it.
What the AI-surfaced practices had in common wasn't backlinks or a clever URL. It was a clear, descriptive name and a real, well-reviewed presence in the place people were asking about.
Honesty requires a split here. Some "factors" are really table stakes — nearly every practice already has them, so they can't explain who wins. Roughly three in four practices already put "dental/dentist" in their name; star ratings clustered between 4.7 and 5.0; external authority barely varied. Those don't get a weight, because there's no spread to learn from. Only the factors with real variation can be ranked:
"Dental/dentist" in the name (~75% of practices have it), star rating (compressed at 4.7–5.0), and external authority (almost everyone at the floor) showed too little spread to rank as predictors. They matter as baseline hygiene; they just can't separate winners here.
Relative weights, scaled to 5. Confident at the extremes (position high; specialty-only keywords slightly negative), provisional in the middle. This is correlational, single-snapshot evidence — not a guaranteed formula.
Across the five cities, the three assistants named 69 distinct practices between them. Only 10 of those were named by all three engines. More than half — 37 — were named by exactly one. The disagreement isn't mostly about ranking; it's about whether a practice makes the list at all.
Returned the longest, most generous lists — 43 mentions across the set — and pulled in more mid-tier and off-roster practices than the others.
With live retrieval, its 40 picks tracked the Maps roster most tightly — the safest proxy for "who's actually ranking right now."
Named the fewest — 28 — and strayed furthest from the visible roster, including names that were hard to reconcile with the local listings.
For a practice, the practical implication is that there is no single "AI" to satisfy. Being named by one engine is common and fragile; being named by all three is rare, and is what durable AEO actually looks like.
The clearest proof that these are two different games: in three of the five cities, an assistant confidently recommended a practice that did not appear in Google's captured top tiers at all. The AI knew about a local, credible practice that our Google roster missed.
None of these had standout citations or authority scores; they were ordinary, well-reviewed local practices with clear names. The lesson cuts both ways — AI can lift a practice that Google's first page buries, and it can ignore one that Google ranks. Neither list contains the other.
The size of the cliff varied a lot by city — and that variation is itself the opportunity. Where the drop from the top ten to the next ten is shallow, AI is willing to look past Google's exact order, and a strong second-tier practice can still get named.
McKinney was the flattest — top-ten and next-ten practices were named at the same 60% rate. That's the cleanest AEO opening in the study: a second-tier practice there has a real shot at the AI shortlist. Naperville was the steepest — 80% of the top ten named, but only 20% of the next ten. There, the gate is everything.
In the smallest markets we studied (Manhattan Beach, Hoboken), Google's roster and the AI lists overlapped most — there simply aren't many practices to disagree about. Overlap is highest where the field is thin; divergence grows with the size of the market.
Top-ten Google Places presence is the single biggest lever for AI visibility. Reviews, profile completeness and relevance still matter — because they're how you cross that threshold.
A town or city name in your practice name was the most consistent marker of practices AI ranked first. Clear, descriptive, geographic naming beats clever branding for findability.
Keyword-stuffed domains and thin directory "citations" didn't move AI visibility. Be a real, well-reviewed local practice with a legible name and footprint.
Being named by one assistant is common and unstable. Track ChatGPT, Gemini and Perplexity separately — and re-check, because the answers drift.
It is a like-for-like comparison, in five real cities, of who Google ranks versus who AI names, with the analysis driven by data we actually pulled rather than impressions.
It isn't a forecast or a formula. Three honest limits: the AI answers are a single point in time and will shift; our Google figure is Places prominence order, not a personalized local pack; and several of our signals (star rating, external authority, "dental" in the name) were so uniform across practices that we can't rank them as predictors — they're baseline hygiene, not differentiators. Where the evidence was thin, we've said so rather than dress it up.
We help dental practices win the new answer-engine game: the Google placement that opens the gate, the naming and presence that get you onto the AI shortlist, and ongoing monitoring across ChatGPT, Gemini and Perplexity — because the answers keep moving.
Find out exactly how today's assistants describe your practice. Request a free AI visibility check — we'll show you which engines name you, where you sit on Google, and the gaps between the two.
Request your free AI visibility checkBecause optimization should be built on evidence, not guesswork. Identity Dental Marketing is a leader in SEO and answer-engine optimization (AEO) for dental practices, and we develop new optimization methods on an ongoing basis. Studies like this one let us ground that work in what actually moves visibility — so the strategies we deploy for clients rest on proven patterns rather than hunches, and the results we deliver are more predictable.
Our research team compares data across every area of the business when building and refining optimization processes — SEO, AEO, Google Ads, and social media marketing alike. We focus on cohesive strategies that produce a synergistic, powerful ROI. Rather than treating each marketing method as a silo, we track each one closely on its own data while valuing how one method can feed and contribute to the effectiveness of another.
Transparency — in our efforts, your rankings, visibility reporting, and results. Organic website traffic is tracked separately from ad-driven traffic, so you have clear visibility into the value of each investment. We hold a strict communication policy and maintain very high standards in both marketing strategy and client support.
Two reasons. First, they're competitive: McKinney, Arlington Heights, Manhattan Beach, Hoboken and Naperville are affluent, dense dental markets where ranking is genuinely hard — the toughest test of any optimization, and where we're transparent about choosing to compete. Second, they match how people actually search. Almost no one searches "dentist in Los Angeles" or "dentist in Chicago"; they search by their city or neighborhood — "dentist in Hoboken," "dentist in Hyde Park," "dentist in Santa Monica." Each of our cities is a bounded local market that behaves like the local searches driving real patient discovery, big metro or not. (We also pulled two whole metros, Los Angeles and Indianapolis, as a contrasting reference — they show why the bare city name is the wrong altitude to compete at.)
Continuously, and not on a fixed schedule. There's run-to-run variation even with identical prompts, day-to-day drift as engines retrieve live results, and larger step-changes when models update. That's exactly why a one-time check isn't enough — visibility needs monitoring.
Largely, yes — and we say so. Google has already baked reviews, relevance and prominence into its ordering, so position works as a composite proxy. That's why it predicts AI visibility so well, and why it's the first thing to fix.
No one honestly can — the answers shift and aren't fully controllable. What the data supports is raising your probability of being named: win the Google gate, name and present your practice clearly, and monitor all three engines over time.