Short answer
Almost certainly, though no vendor publishes the weighting. Reviews are third-party evidence an assistant can read without taking your word for it, and 40% of patients who searched for a provider said review sites influenced them. The rules on obtaining them are federal law, not etiquette.
Key takeaways
- Reviews sit outside your website, which is what makes them useful to a model that is cross-checking you.
- Review sites influenced 40% of patients who searched for a provider (rater8, April 2026, n=714, vendor survey).
- Since 21 October 2024, the FTC rule on consumer reviews makes fake and AI-generated reviews, insider reviews and incentives tied to sentiment enforceable violations, with civil penalties.
- Recency matters as much as volume. Forty reviews that stop two years ago describe a practice that may no longer exist.
Ask an assistant who the best endodontist in a mid-sized city is, and read what it cites. It will rarely be anybody’s home page. It will be directories, review sites and the occasional local article, because those are the sources that describe a practice without the practice writing the description.
Why reviews carry weight a website cannot
Everything on your own site is you describing yourself. A model has no way to verify it and every reason to discount it. Reviews, directory entries and press are independent, which is why they get pulled into answers about who to see. This is the part of the work that cannot be done by writing better pages.
The patient-side numbers support the same conclusion from a different angle. Among the 714 people in rater8’s April 2026 survey who had searched for a provider, 40% said review sites influenced their choice, close behind friends and family at 43% and ahead of AI tools at 36%.
What the law now says about getting them
The FTC finalised a rule on consumer reviews and testimonials in August 2024, effective 21 October 2024. It prohibits fake or AI-generated reviews, reviews by people with no real experience of the business, insider reviews presented as independent, and incentives conditioned on expressing a particular sentiment. Violations carry civil penalties.
| Practice | Status | Why |
|---|---|---|
| Asking every patient, at the same moment in the visit | Fine | No selection by expected sentiment, no incentive |
| A follow-up text with a direct link | Fine | Convenience is not an incentive |
| Offering a discount for any honest review | Risky | Incentives must never be tied to sentiment, and disclosure obligations attach |
| Offering a gift card for a five-star review | Prohibited | Compensation conditioned on sentiment |
| Staff or family posting as patients | Prohibited | Insider reviews presented as independent |
| AI-written reviews from people who were never seen | Prohibited | Named explicitly in the rule |
| Asking only the patients you expect to be happy | Risky | Selection that misrepresents the picture, and easy to prove from records |
Volume, rating or recency: which one to work on
Recency, then volume, then rating. A steady arrival of reviews describes a practice that is currently operating and currently good. A perfect average built three years ago describes history. Nobody outside the platforms can tell you a threshold number, and anyone quoting one precisely is guessing.
A practical target most practices can hold: a handful of new reviews every month, asked for the same way every time, across the two or three platforms your patients actually use. That beats a burst of thirty after a campaign, which looks exactly like what it is.
What to do about the bad ones
Respond briefly, without confirming anyone is a patient. Privacy law applies to your reply as much as to your records, and a defensive response that reveals clinical detail is a much larger problem than the review was. Invite the conversation offline and leave it there.
A visible pattern of calm, non-defensive replies does more for how a practice reads than a spotless average. It also gives an assistant something to describe other than the complaint.
The limits
No platform publishes how its answers weigh reviews, so nobody can promise you a result at a given number. What is certain is the downside: fabricated or incentivised reviews are now a federal enforcement matter, and the same behaviour that triggers a penalty also tends to look artificial to the systems you were trying to influence.
Questions people ask
How many reviews does a practice need before AI notices?
Nobody outside the platforms can answer that, and a specific number should make you suspicious. Aim for steady recent reviews on the two or three platforms your patients use, rather than a threshold. Recency signals that a practice is currently operating.
Can I offer patients a discount for leaving a review?
An incentive tied to sentiment is prohibited under the FTC rule effective October 2024, and any incentive at all creates a disclosure obligation. Asking every patient the same way, with no reward attached, is simpler and carries no risk.
Can I use AI to write review responses?
Responses, yes, with a human reading every word before it posts. Reviews themselves, never: the FTC rule names AI-generated reviews from people who did not experience the business. Responses must also avoid confirming who is a patient.
Should I ask patients to mention a specific treatment?
You can ask people to describe their experience in their own words, and that naturally includes what they came in for. You cannot script it, and you cannot tie anything to what they say. Scripted reviews read as scripted to people and to models.
Do review responses affect AI answers?
There is no published evidence either way, so treat it as unproven. Responses are worth writing because prospective patients read them, and because a calm pattern of replies gives an assistant something to describe besides a complaint.
Sources
- Federal Trade Commission, final rule on consumer reviews and testimonials, 16 CFR Part 465 (announced August 2024, effective 21 October 2024): The prohibitions on fake and AI-generated reviews, insider reviews, and incentives conditioned on sentiment, and that violations carry civil penalties.
- rater8, 2026 Patient Choice Report (April 2026, n=992 US adults, vendor survey): Review sites influenced 40% of the 714 respondents who searched for a provider.
Published . Last reviewed .
Want this in your business?
Book a call and we will scope exactly which part of your operation an agent should take first.
Book a call with the team