PublishedVerifiedEvery 6 weeksSources7 namedAuthored bySquareRank Team
Therapists · AI search · § 4.1.2 · The wedge
Therapist AI Search Visibility on Squarespace
When a prospective client opens ChatGPT and asks "find me a therapist in Austin who takes evening appointments and accepts Aetna", the engine answers with three or four practitioners. Whether one of them is you depends on five things — and none of them are the Psychology Today profile that ranks your name on Google. AI engines pick a different set of pages to cite, and the shape of the win for therapists in 2026 is the shape of the install that ships those specific signals.
This leaf is the wedge across the entire therapist cluster. The intent-shaped query is exactly the kind of question Search Engine Land's 2026 GEO research1 shows AI engines absorb first, and it is exactly the question Psychology Today's templated profile cannot answer well. The install changes for a therapist site is not novel ChatGPT magic — it is the same five-step framework every Pillar 1 page describes, applied to a vertical where intent-shaped queries dominate and entity recognition lives on the practitioner, not the practice.
§01The query shape
The intent-shaped therapist query that AI engines absorb first
Generic 'therapist [city]' queries are Google queries — Psychology Today, group practices, and ad-dense pages own them, and the SEO playbook for them is fixed. The queries that move in AI search are the ones with embedded constraints: 'therapist Austin evening appointments', 'therapist Austin who takes Aetna and specialises in trauma', 'EMDR therapist Denver', 'trauma therapist for first responders in Phoenix'. These queries are too long for Psychology Today's directory filters to answer well, and they are precisely the shape AI engines absorb most aggressively in 2026.
The shift is mechanical. Search Engine Land's 2026 GEO guide1 documents the pattern: AI engines lift answers from passages that match the user's full constraint set, not just the head term. A page that addresses "EMDR therapist who takes Aetna in Denver" with named modality, named insurance, named city, and a real schedule beats a generic "EMDR therapy" article on a major directory because the latter answers the head term but not the constraints. The engines reward the specificity.
The reason this is a therapist wedge — and not a wedge available to every small business — is the density of valid constraint dimensions. A therapy query commonly carries seven constraints stacked: location, modality, insurance, sliding-scale availability, evening/weekend hours, identity/affirming specialty, age group served. Most service-business queries carry two or three. The compounding effect is that the long-tail vocabulary in therapy is much larger than in most niches, and the number of specific pages that can earn a citation for a specific constraint stack is much smaller than it appears.
Why intent-shaped queries move AI citations first
~25%
projected drop in traditional search volume in 2026 as AI engines absorb intent-shaped queries.
What ChatGPT and Perplexity actually cite for therapy queries
ChatGPT and Perplexity cite pages they can reach, pages with passages that answer the user's full constraint stack in 134-167 words, pages with entity-recognised authors, and pages whose schema confirms the entity is a real practitioner. The four filters compound. A therapy page can hit one filter and miss the others, and the citation lands on Psychology Today instead. The job of the AI install is to clear all four filters on the same page, in the order the engines apply them.
The first filter is crawler access. Squarespace's AI exclusion checkbox6 ships unchecked by default, but a meaningful share of therapy sites toggled it on after 2024-era "protect client data" advice — advice that conflated client data (which is never on the public marketing site) with public marketing content (which is what AI engines actually read). The first audit step is verifying the box is still unchecked and that ChatGPT-User, OAI-SearchBot2, and PerplexityBot3 can all reach the relevant pages.
The second filter is passage shape. AI engines prefer pages where the first 134-167 words under each H2 directly answer the section's question without requiring a click — the format Search Engine Land calls "answer-first, expand for context". A therapist page targeting "EMDR therapist Denver" should open with a bolded passage stating, in one or two sentences, that the practice offers EMDR in Denver, with the named practitioner, the insurance accepted, the schedule, and the relevant licensing. Below the lead the page expands with depth.
§03The content
The 134-167 word answer shape that wins therapy citations
A citation-target page for a therapy practice opens each H2 with a bolded one or two sentence lead, between 134 and 167 words, that names the practitioner, the modality, the location, the relevant insurance accepted, and the schedule reality. Below the lead, expand with depth — modality background, what a first session looks like, named cohort experience, state coverage if applicable. The lead is what the engine quotes; the expansion is what the engine reads to confirm the page is real.
The contrast with a Psychology Today profile is the strategic point. A PT profile is field-driven and front-end-filtered, so the prose on the profile is sparse — name, modalities (as tags), insurance (as tags), a short bio, a few specialties (as tags). AI engines can read all of it, but they cannot extract a fluent passage from a tag set. A Squarespace page that puts the same information into a 134-167 word answer-first lead reads as a citable passage and the engine quotes it; the same information distributed across PT's profile fields reads as a database row and the engine cites the directory.
Three details inside the example matter for AI citation. Specific modality names (EMDR, trauma-focused CBT) — not "anxiety" or "depression" as topic tags. Specific insurance accepted by name — not "most major insurers". Specific schedule constraint — "Tuesday and Thursday evenings" not "flexible hours". The engines reward this specificity because the user's query asked for it, and the page that supplies it directly is the page the engine has the highest confidence quoting.
§04Entity
Entity wiring: the practitioner Person + knowsAbout array
For most small businesses, the entity AI engines need to recognise is the brand. For therapy practices, the entity that decides citation is the practitioner. AI engines confidently cite a named human with credentials, public profile links, and a knowsAbout array of real modalities far more readily than they cite an anonymous practice name. Inject Person JSON-LD on the practitioner's bio page with sameAs links to a state board listing or APA roster and a knowsAbout array of named modalities — that single block is the highest-leverage schema change on a therapist site.
The knowsAbout property5 on a Person schema accepts a freeform array of subjects the person is knowledgeable about, and AI engines read it as the canonical list of named modalities and specialisations. For a therapist this is the place to put "Cognitive Behavioral Therapy", "Eye Movement Desensitization and Reprocessing", "Trauma-Focused Cognitive Behavioral Therapy", "Internal Family Systems", "Acceptance and Commitment Therapy", "Dialectical Behavior Therapy" — the real modality names, not topic tags. The medicalSpecialty value on the practice's MedicalBusiness4 covers the regulatory frame; knowsAbout covers the user-facing search vocabulary.
JSON-LDPractitioner Person schema with knowsAbout — paste into the practitioner's bio page Page Settings > Code Injection
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Person","name":"Jordan Lee, LPC","jobTitle":"Licensed Professional Counselor","url":"https://yourpractice.com/team/jordan-lee/","worksFor":{"@type":"MedicalBusiness","name":"Maple Avenue Therapy"},"hasCredential":"Licensed Professional Counselor, Texas, since 2018","knowsAbout": ["Cognitive Behavioral Therapy","Eye Movement Desensitization and Reprocessing","Trauma-Focused Cognitive Behavioral Therapy","Acceptance and Commitment Therapy","First Responder Mental Health"],"sameAs": ["https://www.linkedin.com/in/jordan-lee-lpc","https://www.psychologytoday.com/us/therapists/jordan-lee"]}</script>
The sameAs links matter. AI engines use them to disambiguate one named entity from another, and a practitioner with LinkedIn + a Psychology Today profile + a state board listing is significantly easier to confidently attribute than an anonymous string. The Psychology Today sameAs is a slightly counter-intuitive inclusion — the directory is a competitor — but it functions as a verification link that the practitioner is a real, licensed therapist, which is exactly the confidence signal engines reward.
§05Directory math
Why AI engines tilt the Psychology Today math
Google's local-pack and organic results carry a structural bias toward high-authority directories — Psychology Today, TherapyDen, Zencare, Headway — because directory domain authority is one of the strongest ranking signals in classical SEO. AI engines tilt this differently. ChatGPT and Perplexity weight passage-level extraction over domain authority, and a 1,200-word therapist page on a Squarespace site can be cited above a 200-word PT profile on a query the engine reads as intent-shaped. The directory math that was settled for a decade is unsettled in 2026 for the first time.
The mechanism is the citation-card constraint. When ChatGPT cites three to four sources in an answer to "find me a therapist in Austin who takes evening appointments", the engine picks sources that gave it a quotable passage matching the user's constraints. Psychology Today profiles are too short and too field-driven to supply a multi-constraint passage; the directory's strength on Google (domain authority + structured data) does not transfer cleanly to the AI citation surface. A Squarespace page with a 134-167 word constraint-matching lead beats the PT profile on the citation card for that specific query, even though the PT profile ranks above the practitioner site on Google for the broader head term.
This is not a universal rule. Short queries with little constraint structure ("therapists in Austin") still favour directories on both Google and AI engines. Long-tail constraint-stacked queries favour deep practitioner pages on AI engines. The strategic implication: a 2026 therapist marketing plan does not have to defect from Psychology Today on day one (the directory still earns its keep on the short head queries), but it does have to build owned non-branded discovery on the long-tail intent-shaped queries — because those are the queries growing fastest, and they are the queries the directory cannot win.
§06Measurement
How to measure AI citation for a therapy practice
AI citation measurement for therapists works the same way it does for any small business: a tracking spreadsheet of 10-15 intent-shaped queries, run weekly across ChatGPT, Perplexity, and a Google AI Overviews trigger, with a column for whether the practice appears and a column for which other sources cite alongside it. GA4 referrer data captures only a fraction of AI traffic (most arrives without a referrer), so the manual log is the primary signal. Expect 6-12 weeks for the first visible movement after a full install.
The query list is the leverage. Generic queries ("therapist Austin", "therapy Austin") are easy to track but unlikely to move; intent-shaped queries ("therapist Austin who takes Aetna evening appointments", "EMDR therapist Austin first responders") are harder to track because they require specific phrasing — but those are the queries where citation movement actually shows up. The right list mixes one or two broader queries with eight to twelve constraint-stacked queries the practitioner genuinely wants to be the answer for.
