GEO for therapists is the practice of structuring a counseling practice’s content, clinician credentials, and entity data so ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Gemini, and Microsoft Copilot name that practice when someone asks where to get help. It matters in 2026 because the directories already own the answer: Psychology Today’s own 2026 media kit reports 21.4 million unique visitors a month across 350,000 therapist and treatment center profiles, and every one of those profiles costs the same $29.95 a month, which means the directory sells scale while each individual listing competes against 349,999 others.
The engines lean on that directory layer hard. An ACM Web Science audit of 900 health queries found AI Overviews appeared on 87.7 percent of them, and that only 46.2 percent of the sources cited in those overviews also ranked in the top ten organic results. Translation: your Google ranking and your AI citation are two different scoreboards. Meanwhile the demand is climbing. The Bureau of Labor Statistics projects mental health, behavioral disorder, and substance abuse counselors to grow 19 percent, from roughly 450,000 workers in 2023 to 534,000 by 2033, one of the twenty fastest growing occupations in the country.
Client behavior moved first. The American Psychological Association’s 2026 Chatbots and Mental Health Survey found more than a third of psychologists now have patients using AI as a supplemental source of mental health support, and a JAMA Pediatrics study put nearly one in five adolescents and young adults at having asked an AI chatbot for help when upset or anxious. Those same people ask the same tool who they should see. Right now the answer comes back as Psychology Today, GoodTherapy, TherapyDen, Zocdoc, Alma, Headway, Grow Therapy, SonderMind, or BetterHelp. Almost never a solo LCSW in Charlotte.
Why do AI engines cite Psychology Today instead of your practice?
Because the directory has structured, verified, machine readable data on every clinician and your website usually does not. Psychology Today publishes license type, license number, state, accepted insurance panels, modalities, session fee range, telehealth availability, and populations served as discrete labeled fields on every profile. Your practice site publishes a warm paragraph about creating a safe space.
An engine assembling an answer to “trauma therapist in Raleigh who takes Blue Cross” needs to match four constraints at once: specialty, geography, insurance, and availability. The directory answers all four in parsed fields. Most therapy websites answer none of them in text a model can extract with confidence.
There is a second reason, and it is structural. ChatGPT’s local retrieval leans on Bing Places and directory listings, and Perplexity surfaces a Places panel sourced from business profiles and directory data. If your name, address, and phone number differ by one character between your website, your Google Business Profile, your Psychology Today listing, and your Alma or Headway profile, the engine treats them as separate weakly attested entities and defaults to the aggregator it trusts.
Curious which therapy and counseling queries in your city name a competing practice instead of yours? Get your free AI visibility audit and see exactly which clinicians the engines are recommending and which sources they pull from.
What do prospective clients actually ask AI before booking a therapist?
They ask condition questions, cost and insurance questions, and modality comparison questions, roughly in that order, and they ask them in full sentences. Nobody opens Gemini and types “therapist near me.” They type “what is the difference between EMDR and CBT for PTSD,” then “how much does therapy cost without insurance,” then “how do I find a therapist who takes Aetna in Brooklyn.”
Condition questions dominate the top of the funnel and are currently answered by the National Alliance on Mental Illness, SAMHSA, Mayo Clinic, and Cleveland Clinic. Modality comparison questions carry the highest booking intent, because the person has already decided to start therapy and is choosing an approach. Cost and insurance questions are the last gate before contact, and they are the questions national publishers answer worst, because NAMI will never tell someone what a 50 minute session costs in Denver.
Logistics questions close the loop. Evening availability, telehealth in my state, waitlist, insurance. Telehealth is not a footnote here: post pandemic analyses put telehealth at roughly 37 to 45 percent of mental health visits, far above the low single digits for non mental health care, so state licensure and virtual availability are ranking facts, not fine print.
Which pages does a counseling practice need to earn citations?
Five page types, each built to answer one question class inside the first 40 words. Engines extract passages, not sites. One long “About Our Practice” page that buries fees, specialties, and licensure in the same scroll gets skipped in favor of three pages that each answer one thing cleanly.
1. The condition page
One page per presenting concern you actually treat: generalized anxiety, PTSD, postpartum depression, ADHD in adults, couples conflict, grief. Lead with a plain definitional answer, then cover what treatment involves, how long a course runs, and what the research supports. Reference the American Psychological Association clinical practice guidelines and SAMHSA by name where relevant. Write these as descriptions of the care you provide, not as symptom checklists.
2. The fee and insurance page
This is the highest value page most therapy sites do not have. National reporting puts a typical session between $100 and $250, with out of pocket averages in the $143 to $174 range and insured copays commonly $20 to $50. Publish your own numbers: intake session fee, standard session fee, sliding scale range if you offer one, the panels you are credentialed with, whether you are in network through Alma or Headway, and whether you provide superbills for out of network reimbursement. Every one of those is a retrievable fact.
3. The modality page
EMDR, CBT, DBT, IFS, ACT, EFT for couples, somatic experiencing. Clients search these acronyms after a friend or a podcast names one, and almost nobody explains them in plain language tied to the conditions they suit. Name the modality, describe it in two sentences, state which presentations it is used for, and note where the evidence base is stronger or thinner. Calibrated accuracy earns citations. Overclaiming loses them.
4. The clinician credential page
Every clinician gets a structured bio: degree and granting institution, license type and number, issuing state board, board certifications, years in practice, supervision status, populations served, languages spoken, and states where they hold licensure or PSYPACT authority. Link the state board verification page as reference text. Anonymous or initials only authorship is the fastest way to lose a citation in a health category, which is the same authorship problem we break down in E-E-A-T for AI search.
5. The logistics and access page
Telehealth states, in person office hours, current waitlist status, intake process, session length, cancellation policy, accessibility, and parking. Zocdoc and Google Business Profile fill this vacuum today because practice sites leave it empty.
How should a therapy practice handle schema markup and entity data?
Use Schema.org MedicalBusiness or Psychologist as the practice type, then mark up each clinician with Person schema carrying hasCredential, memberOf, and knowsAbout properties. MedicalBusiness supports medicalSpecialty, availableService, priceRange, and areaServed, which map almost exactly to the four constraints a client query carries.
Add FAQPage schema to your fee page, your modality pages, and your intake page. That single move is what turns a paragraph into an extractable question and answer pair, and it is the mechanic behind most of the citation wins we cover in FAQ content for AI search.
Then do the boring part. Audit your name, address, phone, and license listing across your website, Google Business Profile, Psychology Today, GoodTherapy, TherapyDen, Zocdoc, and any platform profile on Alma, Headway, Grow Therapy, or SonderMind. Headway alone reported more than 70,000 credentialed providers as of July 2026 and Grow Therapy roughly 17,000, so platform profiles are now a major share of the public record about you. Make every one of them agree, character for character.
Does HIPAA limit what a practice can publish for GEO?
HIPAA restricts protected health information, not marketing facts, and almost everything GEO needs is a marketing fact. Fees, licensure, modalities, panels, office hours, telehealth states, and clinician credentials are all practice information, not client information. None of it touches PHI.
Where the rule bites is testimonials and case examples. Most state boards and the APA Ethics Code restrict soliciting testimonials from current clients, and any narrative detailed enough to identify someone is a problem regardless of consent. That creates a real gap: therapy practices cannot build the review volume a med spa or a dental office can, which is exactly why structured credential data and specificity have to carry the trust signal instead. This is the same authority substitution problem covered in GEO for healthcare.
The workaround is not reviews. It is verifiable third party attestation: state board license lookups, board certification registries, professional association memberships, faculty pages, published articles, quoted commentary in press. Engines weigh corroboration across independent sources, and licensure records are among the cleanest corroboration signals in any regulated field.
How do you measure whether AI engines are naming your practice?
Run the same prompt set across ChatGPT, Perplexity, Gemini, Copilot, and Google AI Mode on a fixed schedule, and log which practices and which sources come back. Do not check rankings. Check names.
Variance is the finding most practices miss. An April 2026 test asking ChatGPT, Claude, and Perplexity to recommend an anxiety therapist in Tacoma returned largely different recommendations from each engine, which tells you these systems are not reading one shared index. Winning on Perplexity says nothing about ChatGPT. You need a per engine baseline.
Build a prompt list of 20 to 30 queries mixing condition, modality, insurance, and geography: “best EMDR therapist in Austin,” “couples counselor near me who takes Cigna,” “child therapist in Phoenix with evening availability.” Record the named practices, the cited domains, and whether the engine names any individual provider at all or defers entirely to Psychology Today and Zocdoc. Re-run monthly. The directory share of those answers is your real competitive gap, and it moves when your structured data does.
Frequently asked questions
Is GEO different from SEO for a therapy practice?
Yes. SEO places your page in a ranked list of ten links. GEO gets your practice named inside a single generated answer that may cite three sources. The ACM Web Science health audit found only 46.2 percent of sources cited in AI Overviews also appeared in the organic top ten, so a page can rank well and never get cited. Both matter, but ChatGPT, Perplexity, and Google AI Mode score them separately.
Should I cancel my Psychology Today listing if I invest in GEO?
No. Keep it and fix it. Psychology Today reports more than a million referrals a year and it is one of the primary sources ChatGPT and Perplexity pull from when answering therapist queries, so your profile is training material for the engines whether or not it sends you clients directly. The strategic error is treating a $29.95 listing as your entire digital presence. Treat it as one corroborating record among your website, Google Business Profile, GoodTherapy, TherapyDen, and any Alma or Headway profile.
Will AI engines recommend an individual therapist by name?
Sometimes, and it varies by engine and by how the question is phrased. Direct requests for a named provider often return directory links plus general guidance, while queries with specific constraints like modality, insurance, and city are more likely to surface individual practices. Google AI Overviews applies extra caution to Your Money or Your Life topics and frequently adds a prompt to consult a professional. The practices that do get named are the ones with consistent entity data across Google Business Profile, Psychology Today, and their own site.
Does group practice size affect AI visibility?
It helps, but not the way people expect. A group practice wins because it can publish more distinct clinician credential pages, cover more modalities, and hold licensure in more states, which produces more retrievable facts. A solo practitioner with a fully structured site, a verified Google Business Profile, complete Schema.org Psychologist markup, and five well built condition pages routinely outperforms a ten clinician group with a single generic page and stock photography.
How long does GEO take to show results for a counseling practice?
Expect first citation movement in 60 to 90 days on lower competition condition and modality queries, and four to six months on high volume city plus specialty queries where Psychology Today, Zocdoc, and BetterHelp are entrenched. Schema and NAP consistency fixes propagate fastest because they change how engines resolve your entity. New long form condition pages take longer because they need to be crawled, indexed, and then corroborated against sources like NAMI and SAMHSA.
What is the single biggest GEO mistake therapists make?
Publishing warmth instead of facts. Pages built entirely around tone, with no fee, no license number, no state coverage, no insurance panels, and no modality specifics, give an engine nothing to extract and nothing to verify. The fix is not colder writing, it is adding a structured facts layer under the same voice. That failure pattern and several others show up in common GEO mistakes.
The practices that win are the specific ones
Therapy marketing has spent a decade optimizing for feeling, because feeling is what makes a person pick up the phone. That still holds on the call. It does not hold in the retrieval step that now happens before the call, where a model is matching four hard constraints against a corpus and yours either contains the answer or it does not. Psychology Today wins by default not because it is trusted more but because it is parsed more easily.
The gap closes with specificity. License numbers, panel names, fee ranges, state coverage, modality definitions, and clinician credentials, written as plain text and marked up in Schema.org, are the raw material an engine needs to name you instead of a directory with 350,000 profiles. Warmth converts the client. Structure gets you into the answer where the client is looking.
Want the exact list of therapy queries where AI engines are naming someone else in your market? Request an audit and we will map the prompts, the sources being cited, and the structured data your practice is missing.
Tagged