Patient review generation for dental practices: what Google, Yelp, the FTC and HIPAA allow

Why reviews matter more to AI systems than most practices assume

Every dental practice owner knows reviews matter. Fewer understand why they matter differently to an AI system building a recommendation than they did to a patient scrolling a search results page.

A human reader evaluating a practice's reviews applies judgment — skimming for patterns, discounting an outlier complaint, weighing an old five-star review against a recent three-star one. An AI system performing the same evaluation is working from signals it can actually parse: the star rating, the review count, how recently reviews were posted, and — where the AI system can process review text — the specific content of what patients said.

This creates a subtly different set of incentives than practices have traditionally optimized for. A practice chasing star rating alone might discourage anything less than a five-star review, or stop actively pursuing reviews once it hits a comfortable total. The first instinct is also illegal — selectively suppressing reviews is review gating, which Google's policies prohibit and which the FTC's rule on consumer reviews and testimonials treats as a deceptive practice. The second is simply inert: a review profile that stopped in 2023 gives a system reading it today nothing current to describe. Neither of those claims requires us to know an engine's weighting, and we do not know it.

Four shapes a weak review profile takes

We have not surveyed dental review profiles on the Front Range, and the free scan reads website markup rather than your Google listing, so this is not a finding about your neighbors. These are the four failure modes worth checking your own practice against.

The review count that stopped growing. Many practices accumulated a healthy review count years ago — often during an initial marketing push, a website launch, or a period when a staff member was actively asking patients — and then the flow slowed to a trickle or stopped. The total looks respectable on paper. But the most recent thing anyone has said about the practice in public is years old, which leaves a system answering a question today describing the practice as it used to be.

The single-platform practice. A practice with 150 reviews on Google and essentially nothing on Healthgrades or Zocdoc has built real strength in one place and left a gap everywhere else. AI systems often draw on more than one source to corroborate a recommendation, and a practice that is only verifiable through a single review source is more fragile than one with a presence across several. Yelp is a deliberate exception throughout this article: reviews there are welcome, but you may not ask for them.

The generic five-star review. "Great experience, highly recommend" is a perfectly nice thing for a patient to write, and it does no harm. But it gives an AI system very little to work with when trying to match a specific patient query. A review that says "came in for a same-day emergency crown and they got me in within two hours" is directly useful evidence for an AI system fielding an emergency dental query. Volume of the first type does not substitute for even a modest number of the second.

No process, just hope. The most common pattern of all is simply the absence of any deliberate system — no consistent ask, no easy mechanism, no follow-up. Reviews trickle in from whichever patients happen to feel moved to leave one unprompted, which produces an inconsistent, unpredictable stream rather than the steady one that supports strong AI visibility.

Building a review generation system that actually works

The shift from hoping for reviews to generating them systematically does not require aggressive tactics or anything that risks violating platform review policies. It requires a consistent, low-friction process applied to every patient interaction where a review is a natural fit.

Ask at the right moment. The point right after an appointment — checkout, or a same-day follow-up message — asks while the visit is still fresh, which a request sent days later does not. Ask every patient at that point, not only the ones who seemed pleased: filtering the ask by expected sentiment is review gating, and it is prohibited.

Make the ask specific and easy. A direct link to the practice's Google review page, sent by text or email immediately after the visit, removes friction that a vague "please review us sometime" request does not. The easier the mechanical process, the higher the completion rate.

Encourage detail without scripting it. Asking a patient an open question like "what was most helpful about your visit today?" as part of the request tends to produce more specific, useful reviews than a generic "please leave us a review" ask, without crossing into asking patients to write anything untrue or coached.

Ask on Google and the platforms whose terms permit it — never on Yelp. Rotating requests across Google, Healthgrades and other relevant platforms builds a presence an engine can corroborate from more than one source. Yelp is the exception, and it is not a soft one: its Don't Ask for Reviews policy prohibits soliciting reviews, and its filter demotes reviews it believes were solicited, so asking can leave you worse off than not asking. Never offer anything in exchange for a review on any platform — the FTC's rule on consumer reviews and testimonials (16 CFR Part 465) bans incentivized and fake reviews outright.

Reply to reviews, and keep every reply free of patient information. A professional, non-defensive response to a critical review adds visible content and shows the listing is managed. For a dental practice there is a hard limit on what that reply may contain: under HIPAA you must not confirm that the reviewer is a patient, name a procedure, reference an appointment, or discuss a bill — even when the reviewer disclosed all of it publicly first, and even when the review is unfair. HHS OCR has brought enforcement actions against dental practices for precisely this. The safe reply thanks the person for the feedback, states the practice's general policy, and invites them to call the office. Everything else is said on the phone.

Make it a permanent process, not a campaign. A month-long push to "get more reviews" produces a spike followed by the same decline that created the problem in the first place. Reviews need to flow steadily, which means the request needs to be built into the practice's standard patient workflow rather than run as an occasional initiative.

What we can and cannot tell you about recency

The common claim in this field is that recency beats volume — that 80 reviews from the last year outperform 300 from five years ago. We do not know that, and neither does anyone else outside the engines. No engine publishes a review weighting, our seven Front Range batches recorded which businesses were named rather than why, and the review counts in those batches were proxy estimates rather than measurements.

What can be said without a study: a review profile that stopped years ago describes a practice as it was, and a system answering a question today has nothing current to work with. A review that names a procedure gives an engine something to match against a procedure query; "great experience, highly recommend" does not. Both of those are statements about what the text contains, not about how an algorithm scores it, and they are the reason to keep asking rather than any promised ranking effect.

More on this topic

• GEO for dentists: how to get recommended when patients ask AI for "the best dentist near me"
• Google Business Profile optimization for dental practices in the age of AI answers
• New patient FAQ content: the highest-leverage pages a dental website can build for AI search

Our free AI Visibility Audit at novasapienlabs.com/audit checks the part of this a machine can see from your website: whether your reputation exists on the page in a form an engine can read, and whether you link the profiles where the reviews live. It cannot see your Google Business Profile and it queries no answer engine — your review profile itself you read on the platforms directly.

New to GEO?

Start with the fundamentals, then apply them to your niche.

More in Dentists

See what AI search can read about your practice.

We scanned six Boulder dental and orthodontic practices in September 2026 — scores ran 35 to 79, median 64. The free scan runs the current version of those checks against your site.

Run the free scan AI search for Boulder dentists