The Therapist FAQ Page: How to Build the Version That Gets Cited by AI

A therapy FAQ page — done right — is the most AI-native content format you can publish. ChatGPT, Perplexity, and Google AI Overviews are built around a single interaction: a user asks a question, the engine produces a direct answer. FAQ pages are pre-structured as exactly that pattern: question, then answer. When an AI engine crawls a well-built FAQ page with FAQPage schema, it finds labeled, extraction-ready content that maps directly onto its output format.

Google acknowledged this alignment — somewhat awkwardly — when it removed FAQ rich results from standard search in August 2023, restricting the Q&A accordion in traditional SERPs to government and health authority sites only. But AI platforms haven't followed Google's lead. ChatGPT, Perplexity, Claude, and Google AI Overviews actively extract FAQPage structured data even when those classic rich results no longer appear. The schema still works; it just works for AI search now, not for blue-link results.

This guide covers what makes a therapy FAQ page AI-citation-ready — the right questions, the right phrasing, the right schema implementation, and where FAQ content belongs on a therapy site.

AI engines retrieve content in chunks and generate answers by combining and paraphrasing those chunks. The content that gets cited most reliably is content already decomposed into discrete, answerable units — because that structure maps onto how AI produces output. FAQ pages do this automatically: each question-answer pair is a self-contained retrieval target.

Research on Generative Engine Optimization (GEO) by Aggarwal et al. (Princeton University, Georgia Tech, and IIT Delhi, presented at KDD 2024) found that structured content — including Q&A pairs, statistics, and lists — significantly outperforms unstructured prose for AI citation. A separate analysis in the same research space found that 78% of AI-generated answers use list formats structurally similar to FAQ's question-then-answer pattern (arXiv:2311.09735). FAQ pages are pre-formatted to match what AI engines already produce.

The practical implication: a page with 20 well-phrased therapy FAQ questions and substantive answers gives AI engines 20 separate retrieval targets across 20 different queries. A services page written as flowing prose gives AI engines one. The FAQ page has dramatically more citation surface area — spread across a much wider range of search queries — than any equivalent amount of prose content.

Most therapy FAQ pages answer the questions the therapist wants to answer — not the questions prospective clients are actually asking. The result is pages loaded with logistics:

• What are your office hours?
• Do you accept insurance?
• Where is your office located?
• How long is a session?
• What is your cancellation policy?

These are reasonable things to document. Clients need to know them. But none of these are what someone types into ChatGPT when they're trying to decide whether to try therapy. They're operational questions — the kind a client asks after they've chosen you, not before.

The questions that drive AI citations are the ones prospective clients ask before they commit. Questions like "How do I know if I need therapy?" or "What's the difference between EMDR and CBT for trauma?" or "What actually happens in an anxiety therapy session?" These queries appear in AI engines constantly. When a therapy FAQ page answers them well, with direct, substantive answers, it gets cited in AI responses that reach real prospective clients.

The logistics questions still belong on your FAQ page — they just shouldn't come first, and they shouldn't be the majority of it.

Build your FAQ page in three layers, in this order:

Layer 1: Clinical questions (highest AI citation value)
These are the questions someone asks before they've found a therapist. Phrase them exactly as a client would type them — including informal phrasing:

• How do I know if I need therapy?
• What happens in a first therapy session?
• How long does therapy usually take?
• What's the difference between a therapist, a psychologist, and a psychiatrist?
• Can therapy actually help with anxiety? (or depression, trauma, relationship issues — match your specialties)
• What if I've tried therapy before and it didn't help?
• Is what I say in therapy confidential?
• How do I know if my therapist is the right fit?
• Do I have to talk about my childhood in therapy?
• Can I do therapy online if I'm uncomfortable coming in-person?

Layer 2: Specialty questions (medium AI citation value)
These are specific to your modalities and the populations you serve. If you practice EMDR, include "How does EMDR work for trauma?" If you work with couples, include "When is couples therapy actually recommended?" Specialty questions match the queries AI engines receive from people who are actively researching a specific kind of help — often with more urgency and more intent to book:

• How does [your primary modality] work?
• What does a [your specialty] therapist do differently?
• Is [specific issue] something therapy can help with?
• How many sessions does [your primary treatment] typically take?
• What does [your specialty] therapy feel like in the first few sessions?

Layer 3: Practice logistics (necessary for clients, low AI citation value)
These serve clients who've already found you and need practical answers:

• Do you accept [specific insurance networks]?
• Do you offer telehealth sessions?
• What is your cancellation policy?
• How much does a session cost without insurance?

A well-built therapy FAQ page has 20-25 questions total — roughly 8-10 clinical, 6-8 specialty, 6-8 logistics. The clinical questions appear first. Each answer should be substantive but not sprawling: long enough to stand alone as a complete response, concise enough that the core answer lands in the first two sentences.

FAQPage schema is JSON-LD code that labels your FAQ content for AI and search engines. It tells engines "this content is structured as questions and answers" so they can extract it reliably without parsing unstructured text. The basic structure:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "How do I know if I need therapy?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Common signs include persistent anxiety or sadness that doesn't improve on its own, difficulty managing daily tasks, relationship conflicts that keep recurring, or a sense that something is off but you can't identify what. You don't need to be in crisis to benefit — many people start therapy to build coping skills, work through a transition, or improve relationships before problems become severe."
      }
    }
  ]
}

Key implementation rules:

• One schema entry per question — don't combine multiple questions into a single entry
• Match what visitors see — the schema text value should match the visible answer on the page; don't hide schema-only content
• Verify with Google's Rich Results Test — go to search.google.com/test/rich-results, paste your URL, and confirm FAQPage schema is detected
• Nest with Article schema on blog-style pages — FAQPage + Article compound markup tells AI engines both the content type and the specific Q&A pairs, improving extraction reliability

On WordPress: Rank Math or Yoast SEO Premium can generate FAQPage schema from a visual editor. Confirm the output is rendering by running the Rich Results Test after saving.

On most platforms: Paste the JSON-LD into a code injection block in the page's <head> section. On WebsiteTherapy, FAQPage schema is auto-generated from your FAQ page content — no manual configuration required.

Both. They serve different citation purposes.

A standalone FAQ page (at /faq or /frequently-asked-questions) carries your clinical and general questions. It's the page AI engines cite when someone asks broad therapy questions — "how does therapy work," "is therapy confidential," "how do I find a therapist." Link it from your main navigation so AI crawlers can discover it on every visit.

FAQ sections on individual service pages carry specialty-specific questions. Your EMDR page should answer "how many EMDR sessions does it typically take" and "what does EMDR feel like." Your anxiety therapy page should answer "how does CBT help anxiety" and "how long does anxiety therapy take." These embedded FAQ sections are how AI engines cite you for narrow modality queries, not just broad therapy queries.

Our analysis of how AI engines choose which therapy pages to cite found that practices with dedicated specialty content — individual pages per modality, each with FAQ sections — appeared in AI-generated answers far more often than single-page practices. The citation advantage isn't just from having one well-optimized FAQ page; it comes from having FAQ content distributed across multiple specialty pages, each of which is a separate citation target for different queries.

Think of it as a two-level system: the standalone FAQ page covers broad questions that every therapist should be answering; service-page FAQ sections cover the niche queries that match your specific training and specialty focus.

Question phrasing is where most therapy FAQ pages leave citations on the table. AI engines match user queries to FAQ questions by semantic similarity — the closer your FAQ question is to how a prospective client would actually phrase it, the more likely the match fires and your content gets cited.

Three phrasing principles:

Use informal, conversational phrasing. "How do I know if I need therapy?" outperforms "Is Psychotherapy Appropriate for My Current Situation?" — not because it's simpler, but because it matches how people actually type into ChatGPT. Write the question as your client would phrase it at 11pm when they're finally admitting they might need help, not as it would appear in a clinical intake brochure.

Include your location in location-specific questions. AI engines receive local therapy queries constantly — "EMDR therapist in [City]," "couples therapy near me in [City]." Including your city in a handful of FAQ questions (particularly for specialty questions with local intent) helps AI engines match your FAQ to those local queries. Don't force location into every question; include it where local intent is natural.

Lead answers with the direct response. Each answer should open with the answer itself, then add context. "Yes, EMDR is effective for anxiety — it was originally developed for trauma, but research has found it reduces anxiety symptoms across a range of presentations" works better than an answer that buries "yes" in the third sentence. AI engines extract the beginning of answers when constructing citations. If the answer isn't front-loaded, AI may not capture it cleanly.

This direct-answer principle is the same one that applies across all AI-optimized content — as covered in our guide to GEO for therapists — but it's especially critical in FAQ format, where extraction happens at the individual Q&A pair level.

Update your FAQ page when your practice changes, and when you notice the same question appearing in multiple client interactions. A rule of thumb: if the same question shows up in a client email, an intake form, and a consultation call — it's common enough to add to the public FAQ.

There's also a technical reason to update periodically. OpenAI's crawlers visit websites roughly 1,500 times less frequently than Google's. Anthropic's crawlers visit even less often. Each crawl may be your page's only evaluation for the quarter. A page that's visibly updated — new questions added, answers sharpened — signals freshness without requiring a rewrite. Adding two or three new questions each quarter keeps the page active from a content-currency standpoint.

When you add questions on the page, update the FAQPage schema to match. A question visible in the HTML but absent from the schema means AI engines can read it as prose, but they lose the labeled Q&A structure that makes extraction fast and confident. Every new question added to the page should have a corresponding entry in the JSON-LD.

Starting with a single, well-built FAQ page is the right move for most practices. Twenty well-phrased questions with direct answers and proper FAQPage schema is a meaningful AI citation asset from day one.

Over time, the highest-impact path is building toward what you might call a FAQ system: a standalone FAQ page covering clinical and general questions, plus embedded FAQ sections on each service page. This isn't a two-week project — it's something that grows organically as you add specialty pages and expand your content.

The fastest wins are: (1) audit your existing FAQ page against the three-layer question structure in this guide — most practices can improve citation value significantly just by reordering and replacing the questions already there; (2) add FAQPage schema if it's absent; and (3) move the clinical questions above the logistics questions. Those three steps, applied to an existing FAQ page, cost nothing and can meaningfully increase how often AI engines find and cite your practice.

Ready to see how all of this fits into a complete AI discoverability strategy for your practice? WebsiteTherapy builds FAQPage schema, structured specialty pages, and GEO-optimized content into every site automatically — so your practice shows up in AI-generated answers without maintaining it yourself. See what's included.