Patients and families increasingly ask ChatGPT, Perplexity, Google AI Overviews, and Claude for therapist recommendations. Here’s how behavioral health practices can win at appearing in those recommendations — and why most won’t.
AI Is Already Changing How Patients Find Therapists
A meaningful share of prospective therapy clients in 2026 are asking AI tools — ChatGPT, Perplexity, Google AI Overviews, Claude — for recommendations before they open Google search. The queries range from specific (“best therapists for complex trauma in San Francisco that take Aetna”) to broad (“how do I find a good therapist in my area”). The AI tools respond with synthesized recommendations drawn from public content, reviews, directory listings, and schema-structured data. The practices that get recommended are the ones whose external authority signals are strong enough to surface in the AI’s synthesis. Most therapy practices in 2026 are invisible to this layer and don’t know it.
What AI Tools Actually Look For
1. Google Reviews with Velocity and Recency
Review volume matters; review recency matters more. A practice with 40 recent reviews (last 6 months) at 4.6 stars outranks a practice with 200 old reviews at 4.8 stars in most AI synthesis. Specific-condition mentions in reviews (e.g., “helped me with my ADHD,” “changed my postpartum anxiety”) compound because they match the specific conditions AI tools pattern-match on.
2. Directory Listings with NAP Consistency
Psychology Today, Healthgrades, Zocdoc, TherapyDen, Inclusive Therapists, and culturally-specific directories like Therapy for Black Girls. AI tools cross-reference directory data. Inconsistent name/address/phone across directories reduces the AI’s confidence in the practice as a signal. Audit quarterly.
3. External Citations in Journalism and Healthcare Publications
A backlink from local press, a quoted expert role in a healthcare article, a guest post on a mental health publication — these are the authority signals AI tools weight heavily because they cannot be gamed as easily as on-page content. Two to four PR placements per year compounds meaningfully over 12 to 24 months. Covered in detail in our PR and SEO article.
4. Schema-Structured Content
LocalBusiness schema for the practice, Physician schema for each clinician, FAQPage schema on condition pages, Person schema for published authors. Structured data makes it easier for AI tools to extract practice facts (name, address, specialties, credentials, fees) accurately. Schema is underused in therapy SEO; it’s a competitive advantage right now.
5. YMYL-Compliant Clinical Content
AI tools weight E-E-A-T signals (Experience, Expertise, Authoritativeness, Trustworthiness) for health-related queries. Clinician-authored content with credentials, peer-reviewed citations, and condition-specific depth outperforms generic blog content. Covered in our content marketing guide.
Good SEO = Good GEO (Mostly)
The signals that drive traditional SEO overlap substantially with the signals that drive AI recommendations — reviews, citations, authority, structured data. The difference is weighting: AI tools lean more heavily on external authority signals (reviews, PR, directory consistency) than Google organic ranking does. Content volume matters less to AI recommendations than it does to organic ranking; content quality and clinician credentialing matter more.
Practices investing in SEO correctly in 2026 are also investing in GEO correctly. Practices that built content farms hoping to rank organically often find those farms don’t get cited by AI tools — which weight authority over volume.
Explore the Top 10 anxiety therapists in Naperville
Why Most Practices Will Be Invisible to AI
The therapy practices that will be invisible to AI recommendations in 2026 share a profile: thin Google review presence, inconsistent NAP across directories, zero PR or external citation work, no schema markup on the practice website, and content written by marketing copywriters without clinical input. The invisibility is quiet — the practice owner doesn’t know they’re being filtered out of AI recommendations because the channel doesn’t show in their analytics. Traffic from AI tools arrives at the website as direct or unattributed organic.
The correction is not complicated; it’s just work that compounds over 6 to 12 months. Review velocity, citation audit, quarterly PR outreach, schema implementation, clinician-authored content. Practices that start now show up in AI recommendations by late 2026; practices that wait a year will find the authority gap harder to close.
Frequently Asked Questions
What is GEO and AIO for therapists?
Generative Engine Optimization / AI Optimization — structuring therapy practice content to be surfaced in AI-generated recommendations from ChatGPT, Perplexity, Google AI Overviews, Claude.
Do AI tools actually recommend therapists?
Yes, increasingly. Smaller volume than Google search but growing; converts at higher rates because patients arrive pre-qualified by the AI’s recommendation.
What signals do AI tools use?
Reviews (volume + velocity), directory NAP consistency, external citations in journalism, schema markup, YMYL-compliant clinical content.
How do I optimize for AI search?
Review velocity, NAP audit, 2–4 PR placements/year, comprehensive schema markup, clinician-authored YMYL content. Compounds over 6–12 months.
For more guides like this, browse the full VYDC blog.