How ChatGPT chooses which med spa to cite

A patient in Scottsdale asks ChatGPT: "Best Daxxify injector for natural-looking results near Paradise Valley?" ChatGPT names two providers. Two of approximately 80 operating med spas in the Scottsdale metro. The other 78 are invisible to that prospect.

The med spa citation mechanics differ from the GLP-1 clinic pattern in one structural way: the injector entity matters as much as the spa entity. Patients pick the injector, not the building. The structured data layer must reflect this.

Signal 1: The injector Person schema

Short answer. Every injector at the spa must be declared as a Person entity with medicalSpecialty, hasCredential array (NP, PA, RN license type), memberOf (state nursing association, ASCN, AAFE, IAPAM), and worksFor linking to the med spa MedicalClinic. Spas with named, credentialed injector schema cite at 3-5x the rate of spas that hide injector identity behind a generic "treatments" page.

Of the 30 med spas audited across Scottsdale, Beverly Hills, and Miami, only 11 declared Person entities for each injector. Those 11 cited at 8.4 queries each on average. The 19 with generic injector pages cited at 1.7 queries each.

Signal 2: Supervising physician in collaborative-practice states

Short answer. NP-led and PA-led med spas in collaborative-practice states (Texas, California, New York, others) must declare the supervising physician as a separate Physician entity with affiliation making the supervisory relationship explicit. Spas that hide the supervising physician trigger downranking on the YMYL trust filter.

This is the single most-common compliance schema gap KailxLabs sees in med spa audits. Spas that map the supervising physician correctly cite materially higher than spas that obscure the supervisory relationship.

Signal 3: Drug entity per injectable brand

Short answer. Botox is a brand. Neurotoxin is the category. Patients ask both interchangeably. Every neurotoxin (Botox, Dysport, Daxxify, Xeomin, Jeuveau), every HA filler (Juvederm, Restylane, Belotero), every biostimulator (Sculptra, Radiesse), every skin booster (Skinvive, Profhilo) needs its own Drug entity. Spas with full mapping cite for both branded and category queries simultaneously.

Of the 11 spas with full injector schema, 7 also declared Drug entities per brand. Those 7 cited at 11.2 queries each on average. The 4 with injector schema but no Drug entities cited at 5.6 queries each.

Signal 4: Before-after ImageObject with injector credit

Short answer. Med spa galleries are the highest-traffic asset on every site. Without schema, the gallery generates traffic but no AI citation extraction. Every gallery image declared with ImageObject (caption, contentUrl, creditText to the injector, dateCreated, subjectOf linking to MedicalProcedure) becomes a citation surface. The injector credit specifically is a corroboration anchor.

Of the 7 spas with Drug entity mapping, only 3 declared ImageObject schema on the gallery. Those 3 cited at 13.7 queries each. The 4 with gallery but no ImageObject schema cited at 8.7 queries each.

Signal 5: llms.txt with med-spa-specific Q&A pairs

Short answer. A well-structured llms.txt for a med spa includes Q&A pairs covering injector credentials, Botox vs Daxxify comparison, filler longevity expectations, biostimulator protocol differences, membership program structure, and state compliance posture. The vertical-specific Q&A is the citation extraction unit.

All 3 spas with ImageObject mapping also had llms.txt at root. Citation share at this tier averaged 14.0 queries each, near the citation ceiling for the test set.

The membership program signal

Beyond the five primary signals, membership programs (Alle, Aspire, in-house) deserve Offer schema with recurring UnitPriceSpecification. Spas with structured membership Offer cite for "med spa membership programs in [city]" comparison queries that competitors hide behind prose. This is a secondary signal worth 1-2 additional cited queries per spa on average.

The Instagram-to-AI search bridge

Med spa prospects often discover providers on Instagram first then verify through AI search before booking. The injector entity schema is what makes that verification step succeed. A prospect who sees a Daxxify result on Instagram then asks ChatGPT to verify the injector reaches the spa only if the injector schema exists.

This is why the injector entity layer matters disproportionately in med spa vs other verticals. The cross-channel verification pattern routes through structured data the spa controls.

Citations

40 US Clinic AI Visibility Audit (Q1 2026)