How AI assistants use reviews when recommending a practice

Updated November 14, 2026 · 5 min read
Short answer

AI assistants do not just count stars. Google says Maps uses Gemini to summarize what people like about a place from its reviews, and assistants that search the web can draw on the review sites and profiles they find. The themes patients repeat, such as long waits or a doctor who explains things well, can become the one line an assistant says about you.

What is documented, and what is not

It helps to separate what the companies have said from what is assumed.

Documented by Google: Google has announced that Maps uses Gemini to give review summaries, so people can find out what others like about a place, grounded in its data about places and insights from the Maps community. Google's local ranking help says prominence, one of the main factors in local results, draws on information such as how many reviews a business has, and that more reviews and positive ratings can help. Google also says Business Profiles may show reviews from other websites.

Not published in detail by anyone: exactly how much weight any assistant gives reviews compared with other signals, which review sites each one reads, or how often they refresh what they know.

So be skeptical of anyone who claims to know the formula. What you can do is test what the assistants say about you today and improve the inputs you control.

Why the words in reviews matter more than ever

A star rating is one number. A summary is a sentence, and it is built from words. When an assistant or a Maps summary describes your practice, it draws on what reviewers actually wrote.

That changes what a review campaign is for. Ten five star reviews that say only "great" give a summary little to work with. Reviews that describe a short wait, a doctor who took time to explain a diagnosis, or a front desk that sorted out an insurance problem give it something specific and favorable to say.

It also means recurring complaints carry more weight. If a dozen reviews over two years mention trouble reaching anyone by phone, a summary may well mention it, even if every other review is glowing. The fix is operational, which is why the guide on why reviews decide which doctor a patient picks starts with tallying your themes.

See your Surge Score in about a minute

Does ChatGPT recommend your practice?

The Surge Score measures five things about your practice website: how often ChatGPT, Gemini and Perplexity recommend you, your PageSpeed, your structured data, your indexing and your reviews, each with its evidence and date. Free.

Ask the assistants yourself: a 15 minute check

Run this once a quarter and keep the answers.

  1. In ChatGPT, Gemini and Perplexity, ask: "Who is a good [your specialty] near [your town]?" Note whether you are named and what reasons are given.
  2. Ask: "What do patients say about [your practice name]?" Note the themes the assistant reports.
  3. Ask the same about each provider by name.
  4. Where the assistant cites sources, note which sites it used: your website, Google, Healthgrades, Zocdoc, Yelp or others.
  5. On Google Maps, open your profile and read any review summary shown.
  6. Compare the themes with your own review tally. Do the assistants reflect the practice as it is now?

Answers vary by wording, location and account, so look for patterns rather than single responses. The guide on checking what AI says about your practice goes deeper.

Shape the themes by changing the experience

It is tempting to try to steer what reviewers write. Do not. Google's policy says merchants should not request that specific content be included in reviews, and that asking staff to solicit reviews that include specific content, including content naming a staff member, is not allowed.

The legitimate path is slower and works better:

  • Fix the operational causes of recurring complaints, such as phone access, waits and billing clarity
  • Ask every patient for a review, so ordinary good experiences are on the record
  • Reply calmly to critical reviews, without disclosing anything about the reviewer
  • Give recent reviews time to outweigh older ones

As the experience improves, the words in new reviews change, and the summaries built from them follow.

Consistency across sources

Assistants that search the web may find your practice on Google, Healthgrades, Zocdoc, Yelp, insurer directories and your own website. If those sources disagree about your address, your providers or what you treat, an assistant has to guess, and it may hedge or repeat the wrong detail.

The same applies to reviews. A practice with a strong Google profile but a neglected Healthgrades page full of old complaints presents a mixed picture to any system reading both.

Claim the profiles that appear for your names, correct the facts, reply to reviews on each one with the same care, and make your own website the clearest, most complete source of truth about your practice. For a platform by platform view, see Google versus Healthgrades versus Zocdoc versus Yelp.

Fake reviews are a worse bet with AI in the loop

Some practices wonder whether a few planted reviews could tilt what an assistant says. Beyond being prohibited, it is a poor bet.

Google says it uses AI to help detect fake reviews, that it can now revisit reviews to find abuse patterns months after they were posted, and that it blocked or removed more than 240 million policy violating reviews in 2024. It also says it shows warnings in some cases where it has recently removed suspicious five star reviews from a place.

A warning on your profile, or a sudden disappearance of reviews, is exactly the kind of signal that could end up shaping how your practice is described. The durable approach is the honest one: real reviews from every patient, a better experience over time, and consistent information everywhere.

Checklist

  • Ask three AI assistants who they recommend for your specialty and town
  • Ask what patients say about your practice and each provider
  • Record which sources the assistants cite
  • Read your Google Maps review summary, if one is shown
  • Fix the operational cause behind any theme you do not want repeated
  • Never ask patients to include specific words in reviews
  • Repeat the check every quarter and compare

Frequently asked

Does ChatGPT read our Google reviews?

We could not find a public, detailed statement from OpenAI on exactly which review sources ChatGPT uses for local recommendations. Assistants that search the web can draw on whatever pages they find, which may include review sites and directories. Test it yourself by asking what patients say about your practice and checking the sources it cites.

Can we correct what an AI assistant says about our practice?

Not directly. You can improve what it reads: accurate, consistent information on your site and profiles, and a steady flow of genuine reviews that reflect your current care. Changes take time to show, so recheck quarterly rather than daily.

Will a few bad reviews dominate an AI summary?

They can if they describe a recurring theme and there are few recent reviews to balance them. A larger base of recent, specific reviews makes any single complaint less likely to define the summary.

Do our replies to reviews affect AI summaries?

No assistant has published how replies are used. Replies are public text, and patients certainly read them, so a calm, privacy safe reply is worth posting for human readers regardless.

Sources

  1. Google Maps updates: Gemini arrives, Immersive View expands and more, Google
  2. Tips to improve your local ranking on Google, Google Business Profile Help
  3. About third-party reviews, Google Business Profile Help
  4. Prohibited and restricted content, Maps user generated content policy, Google Maps Help
  5. Google Maps uses AI to fight fake Business Profiles, Google
Prefer to talk it through?

Book a strategy call with the team.

Twenty minutes. We'll walk through the specific opportunities in your market and what a MedPracticeGrowth engagement would look like for your practice.

AI patient discovery audit

Is AI recommending you, or your competitor?

See who ChatGPT, Gemini and Perplexity recommend to patients in your area.

Free. Built by physicians.