Updated 3 October 2026 · Therapist Website Studio
How AI search reads a therapy website
What can an answer engine quote, and what can it not invent for you?
It can only use public pages it is able to retrieve. A useful therapy site states the service, the place or online scope, the fee and the way to enquire on those pages. That supports Google’s AI features and other answer engines. It does not guarantee a mention.
A worked example
Imagine someone asks which facts a Manchester counselling page should make obvious. On the Morrow Lane example, individual counselling, anxiety, grief, the Manchester room, online UK sessions and the £70 fee each have a place to live. A real practice uses its own facts in those same places.
Google’s current guidance treats this as part of SEO: useful original pages that can be crawled. It does not ask you to write for a model, and it does not require a special file to be included.
Different systems fetch in different ways
Googlebot crawls for Search, which AI Overviews and AI Mode draw on. Google-Extended is a robots.txt control for Gemini training and grounding, not a separate crawler and not a Search switch. ChatGPT search uses OAI-SearchBot, Claude search uses Claude-SearchBot, and Perplexity uses PerplexityBot. Training crawlers and the fetches a person triggers are separate. Naming them is how the page is planned. Blocking or allowing them is a choice you make in robots.txt, not a ranking product we sell.
Structured data should repeat the visible facts. It should not add a fee, a title or a review that the visitor cannot see.
Sources
Ready for a clear next step?
Request a free website design review if you already have a site, or book a free 20-minute consultation if you are starting from scratch.
Or email hello@therapistwebsitestudio.co.uk.