AI Search for Therapists: Discoverability in ChatGPT, Claude, Perplexity, Gemini, and Grok

Between 2023 and 2026, the search landscape for care and therapy has fundamentally changed. Where people formerly opened Google first, they now increasingly open ChatGPT, Claude, Perplexity, Grok, or Google's own AI-Overview. The question "which therapist suits me" is no longer answered with a list of blue links, but with a summarizing answer that cites a handful of websites. If you are not among that handful, you do not exist to that searcher.

How AI Search Engines Decide Who to Cite

Unlike Google's classic ranking, an AI search engine is a language model that not only indexes the webpage but also reads its content. It looks for three things: a clear answer to the posed question, a recognizable source with expertise, and a structure in which it can unequivocally cite the information. Pages with long introductions, marketing jargon, or vague claims are skipped.

Six Optimizations That Have Immediate Impact

1. Answer Upfront, Context After

Start each page with a block of 40 to 80 words that answers the core question in plain language. Our TL;DR blocks (like at the top of this article) are built precisely for this. AI models scan the first paragraph to assess whether the page truly answers the question.

2. Question-and-Answer Structure with FAQPage Schema

A visible FAQ section can answer questions in compact, self-contained blocks. FAQPage JSON-LD may help machines understand the question-and-answer relationship, but Google no longer shows FAQ rich results for most sites and the markup does not guarantee an AI citation.

3. Transparent Author Information

At the bottom of each article, state who the author is, what education or registration that person has, and when the article was last updated. AI models heavily weigh E-E-A-T signals, especially for health content. An article without a visible author is rarely cited.

4. Source Attributions to Authority

Refer in your text to authoritative sources: the Polyvagal Institute, the Dutch Association for Psychotherapy, Schema.org, or Google's own E-E-A-T guidelines. Not to play SEO games, but because AI models judge pages with external references as more reliable.

5. Clear Heading Structure with Semantic H2s

Give each section an H2 that unambiguously names the question or topic, and give that H2 a stable anchorId. AI models prefer to cite text blocks they can precisely pinpoint, not entire pages. A well-structured article will be cited by sections, a long story without headings will not.

6. Unambiguous Facts, No Marketing Claims

Avoid phrases like "the best therapy for trauma" or "fast results." AI models filter out superlatives and unsubstantiated claims. Instead, write: "trauma-oriented therapy where an average of 8 to 15 sessions are needed for noticeable change." A verifiable claim lands; a marketing claim is skipped.

What Each AI Search Engine Does Differently

The five major AI search engines work slightly differently. ChatGPT (OpenAI) has used built-in web search since 2024 and cites sources inline; pages with clear TL;DR blocks and FAQPage schema land here most often. Claude (Anthropic) has had web search since 2025 via its own index plus direct web calls, and heavily weighs readability, length, and the absence of marketing jargon; long nuanced explanations in plain language work better here than short SEO text. Perplexity is strong in citation transparency and always shows the sources used under each answer; it rewards authority signals and external source attributions. Google Gemini (and AI Overviews) relies heavily on existing Google ranking plus E-E-A-T, so classic SEO remains the foundation here. Grok (xAI), in addition to the open web, also draws directly from real-time signals from X (formerly Twitter); pages with an active author and recent updates are weighed extra heavily here, and very recent topics are cited more often than in other models.

Practically speaking, you don't need to create five versions of your content. The six optimizations above work for all five. The difference lies in which signals each engine weighs extra heavily: schema for ChatGPT, readable depth for Claude, external sources for Perplexity, classic authority for Gemini, and recency plus a visible author for Grok.

What This Means for Your Site

Many of these optimizations are the same ones that already make your site stronger for classic SEO and for human readers. The beauty of AI search is that it rewards honest pages: clear, structured, signed by a real expert. Those who do their work well are cited. Those who try to mislead are skipped.

Next Steps

Start with the page you consider most important (often your methodology or a main complaint page) and add a TL;DR block at the top, plus an FAQ section at the bottom with four questions your clients ask you during an intake. Then test a query in Perplexity, Claude, ChatGPT, and Grok about your specialty in your city, and see if you are cited. If not, you know where the next improvement step lies.