Medspa AI Visibility Tracking: Tools, Metrics, and a 30-Day Test

Can a tool show whether AI recommends your medspa?

Medspa AI visibility tracking is the repeatable observation of whether selected AI systems mention, cite, or recommend a clinic for defined prompts and locations. Tools can show those observations. They cannot prove a universal recommendation rank. The result depends on the platform, prompt wording, location, date, account context, and the competitors included in the test.

Start with the evidence the platforms expose. Google's generative AI performance report shows impressions for links to your site in eligible Google AI features. Microsoft's AI Performance report shows citations, cited pages, and grounding queries from Microsoft Copilots and partners. Neither report tells you every time an assistant named an unlinked clinic or whether a recommendation produced an appointment.

For competitor comparison, add a local prompt tracker or documented manual test. Keep the test narrow enough to repeat; a large dashboard is weak evidence when its prompts do not match the clinic's services, locations, or patient decisions.

What should a medspa measure in AI answers?

Use a measurement ladder because each label answers a different question. A mention is the appearance of the clinic name. A recommendation is the presentation of that clinic as an option for the request. A citation is a source link attached to the answer. Share of voice compares appearances within a defined prompt and competitor set. A referral visit records a website arrival from an identifiable AI surface. A clinic outcome requires a separate inquiry, booking, or accounting record. Each higher layer needs evidence from the systems below it. A citation can support a visibility claim, but it cannot support a lead or revenue claim by itself. Keep the full answer behind the aggregate number so a neutral source citation is not mislabeled as an endorsement.

Use the 100-point Recommendation Evidence Score

Score the test design before buying a yearly plan or reporting a win. The original Recommendation Evidence Score gives the most weight to reproducible captures and mention or citation evidence. Reproducibility matters more than dashboard sophistication because consistent documentation prevents misinterpretation and attribution mistakes. Prompt coverage and local controls matter too, but neither can rescue a report that hides the underlying answer.

Download the 48-control AI visibility tracking scorecard. Give credit only when the named evidence exists. A score below 70 is a planning aid, not a reporting system. A score from 70 to 84 can support directional monitoring. Require 85 or more before using the result in an executive report, and preserve the export or screenshots behind every claim.

Which AI visibility tools fit a local medspa?

There is no universal best tool. First-party reports answer what happened to the clinic's own website. Paid trackers answer broader prompt and competitor questions, but their datasets, locations, update schedules, and scoring formulas differ. Choose the smallest system that answers the clinic's actual decision.

Prices were verified on November 4, 2026. They can change, so confirm current terms, limits, and cancellation rules before purchase.

What do Google and Bing reports actually tell you?

Google rolled its generative AI performance report to websites worldwide on August 31, 2026. It counts impressions when a link to the site is shown in supported generative AI features and breaks the data down by page, country, device, and date. Google warns that the report may not appear when a property has too few impressions. Microsoft's AI Performance report counts citations from Microsoft Copilots and partners. In June 2026, Microsoft added Intents, Topics, Citation Share, and Compare. The current interface labels results as sampled and says intent and topic labels are AI generated. These reports provide site-owned evidence, not a complete prompt log or competitor leaderboard. Preserve their definitions and limitations when the numbers move into an owner report.

Use both reports as source-side evidence. If a page earns impressions or citations, inspect the cited page, the query or topic detail available, and the time trend. Then look for referral visits and site actions separately. The medspa AI search optimization guide covers the crawl, evidence, local fact, and page improvements that precede measurement.

How should a medspa build its prompt set?

Start with 20 to 30 prompts, not hundreds. Split them across four jobs: finding a service, comparing nearby clinics, understanding price or process, and checking trust or convenience factors. Every prompt needs an owner, location, intent, service, and reason for inclusion.

Keep branded and nonbranded prompts separate. "Does [clinic] offer [service]?" measures fact retrieval. "Medspas near [neighborhood] offering [service]" measures discovery. "[Clinic A] vs [Clinic B] for online consultation requests" measures a comparison. Combining the three into one score hides what changed.

• Use only services, locations, policies, and provider facts the clinic currently approves.
• Exclude prompts that ask the assistant to diagnose, select a treatment, promise results, or replace a qualified clinician.
• Keep spelling variants and conversational rewrites in the same intent group instead of treating each as a separate market.
• Freeze the baseline prompt list for 30 days. Document any correction instead of silently editing history.

Why do location and repetition change the answer?

Local AI results are not fixed rankings. A useful test records the place, repeats the prompt, and treats variation as part of the result. One answer from one account at one address is a sample, not the clinic's standing across a market.

BrightLocal's local AI visibility study tested 200,000 nonbranded prompts and 1.9 million citations. The vendor repeated each prompt ten times at each grid point and reported that ChatGPT recommendation overlap between two points in the same city was 36.2%, while the two tested Google surfaces were around 47%. This is vendor research, not a universal benchmark, but it demonstrates why a local test needs more than one point and one run. For a practical clinic baseline, choose the clinic address plus two to four meaningful patient areas. Repeat priority prompts at least three times under the same conditions. Report the response spread, not only the best result, and do not use location simulation to imply what every real person will see.

How should competitor comparison work?

Define competitors before reading the score. Include clinics that share the relevant service, patient area, and business category. For example, a national directory, hospital department, product brand, and independent medspa may all appear in answers, but they are not interchangeable competitors. BrightLocal says its competitor comparison auto-detects local competitors from location and category, then compares visibility score, mention rate, share of voice, average position, and platform results. Automatic discovery is useful for finding unexpected entities. The clinic should still approve the comparison set and document additions, removals, mergers, name variants, and closed locations.

Use relative metrics as observation aids. State the formula and denominator beside every share-of-voice figure. Ahrefs, for example, warns in its AI visibility metrics documentation that brand and competitor entity setup can materially change AI share of voice. A score without its entities and prompts is not reproducible.

How do you connect visibility to website and clinic outcomes?

Build the handoff before the dashboard. Tag AI referral traffic where the platform provides a source, record the landing page and meaningful site action, and confirm whether the inquiry reached the clinic's approved system. Keep direct, organic, referral, and unknown attribution available instead of forcing every unclear visit into an AI bucket. The landing page still has to earn the next step. Match the answer's service and location context, keep provider and policy facts current, and give the visitor one clear route to ask a question or request a consultation. The medspa conversion path guide maps the evidence from search result to confirmed appointment. The parent medspa website optimization service explains how TheClinify maintains the underlying pages and booking paths.

Do not report a recommendation, citation, or session as a patient outcome. A qualified inquiry requires clinic-side receipt and status. An appointment requires the booking record. Revenue needs accounting evidence and a written attribution rule. Even then, use "associated with" unless the method can support a stronger causal claim.

Run this 30-day medspa AI visibility test

Use the first month to establish a defensible baseline, not to chase daily score changes. The goal is a repeatable evidence packet and a short improvement backlog.

What limitations belong in every AI visibility report?

State that AI answers vary, platform datasets differ, and prompt trackers observe selected conditions rather than every user response. Note whether results came from a first-party report, a vendor's tracked prompt set, a manual capture, analytics, or a clinic record. Do not use visibility data to make clinical claims, assess individual suitability, or imply care quality. A recommendation is generated text, not a professional referral or credential review. Route treatment and health questions to qualified clinic staff under the clinic's approved process. Last verified November 4, 2026. The editorial policy explains source review, corrections, and update dates. The About page identifies the publisher and its scope.