Imagine a fairly ordinary moment. Someone in Franklin has been considering Botox for a few months. She has never booked an injectable treatment, does not know which credential matters, and would prefer a practice close to Cool Springs. A few years ago, she might have typed “Botox near me,” opened several tabs, and worked her way through the map results.
She can still do that. She can also ask Gemini or ChatGPT a much fuller question: which med spas near Cool Springs are known for natural-looking results, have experienced injectors, and make first-time patients feel comfortable? The response may arrive as a short list, with reasons. That list becomes the beginning of her consideration set.
For the practice owner, the change is easy to underestimate. Search did not disappear. It became more interpretive. The system is no longer being asked only to retrieve nearby pages. It is being asked to understand the treatment, the place, the provider, the reputation, and the patient’s uncertainty—then make a recommendation.
A recommendation is assembled from more than the website.
The practice website is still important because it is usually the clearest first-party statement of who the business is and what it offers. A good treatment page can explain candidacy, process, risks, recovery, provider experience, and how to book. A good location page can connect the practice to the right city and neighborhood. Structured data and clean technical foundations make those facts easier to retrieve.
But the website is the business talking about itself. Recommendation systems often look for support elsewhere: a Google Business Profile, reviews, professional credentials, local directories, association pages, news coverage, interviews, social profiles, and other sites that independently connect the same business to the same treatments and location.
The system is trying to answer two questions at once: “Is this relevant?” and “Is there enough evidence to say so with confidence?”
That is why two practices with equally polished websites can produce different answers. One may have a stronger location trail. Another may have clearer provider credentials. A third may be mentioned by the AI system but missing from the local pack because the evidence that connects its exact location to the requested treatment is weaker or less consistent.
We saw this happen in Franklin.
Frontier Rank tested three Franklin medical spas using four fixed buyer questions, one market, and one scoring model. The questions covered Botox, fillers, reputation, and personalized injectable care near Cool Springs. All three websites received the same readiness score: 74 out of 100.
The discovery results were not equal. OVME Cool Springs produced the strongest composite visibility across the measured evidence classes. Glow Medical Spa appeared in two of the three Google and Maps checks and held the best observed average position when present. Grace Aesthetics was mentioned in both AI checks and ranked second in the returned Yelp set, but did not appear in the three matched Google and Maps checks.
This was not a declaration that one practice is better than another. It was a dated snapshot of what the systems could retrieve and support at that moment. The useful finding was the disagreement itself. A website score did not explain it. A single screenshot would not have explained it either.
Most practices do not need more content. They need better evidence.
“Publish more” is convenient advice because it can be repeated without diagnosing the problem. It is not always useful. If the business name, address, hours, category, and booking path disagree across public profiles, another article will not repair the identity problem. If the Botox page never names the location or explains who provides the treatment, a broad lifestyle post will not make that page a better answer. If the practice has little credible support outside its own website, rewriting the homepage may leave the real gap untouched.
The better sequence is to measure first. Lock the exact practice location and a small group of commercially important treatments. Write down the questions a patient might realistically ask. Record which businesses are mentioned, how they are described, what position they hold when position is available, and which sources appear to support the answer.
Only then should the practice decide what to change. Sometimes the highest-value work is a stronger treatment page. Sometimes it is cleaning up profiles and citations. Sometimes it is helping excellent clinicians become visible through bios, credentials, interviews, or association listings. Often it is a combination, but the order should come from the observed gap.
The med spas that learn fastest will keep the questions stable.
AI answers change. That is not a reason to stop measuring; it is a reason to measure carefully. A practice cannot claim improvement because a different prompt produced a friendlier result three weeks later. The question, location, business entity, provider set, and scoring rules should remain stable enough to make the comparison meaningful.
A useful operating record is surprisingly plain. It shows the original question, the original answer, the competitors that appeared, the cited or supporting sources, the changes made by the practice, and the answer returned during the retest. Wins belong in the record. So do non-wins and provider failures.
Over time, that record changes how marketing decisions are made. The practice stops asking whether it “did SEO” or “posted enough.” It can ask whether more patient questions now lead to the business, whether the recommendation is stronger, whether the supporting source trail improved, and whether those changes correspond with qualified calls and bookings.
What should a med spa owner do next?
Start with five questions, not fifty pages. Choose a location, two or three important treatments, one reputation question, and one question that reflects what makes the practice meaningfully different. Run them across the search and answer surfaces that patients are likely to use. Save the results.
Then look for the repeated pattern. If the same competitors appear everywhere, study the evidence around them. If your practice appears in Maps but not in AI answers, examine whether the treatment and provider information is clear enough to support a direct recommendation. If AI systems mention the practice but local results do not, verify the location trail, Google Business Profile, citations, and treatment-to-location connection.
The point is not to chase every answer engine. It is to make the business easier to understand and easier to trust across the places that shape a patient’s shortlist. That is durable work, even as the interfaces change.
See what a patient sees.
We can build a private snapshot showing who appears instead and which evidence gaps may be influencing the result.