A growing share of patients now describe their problem to an assistant rather than typing keywords into a search box. The answer they get names a small number of clinics, and the criteria for being one of them are not the same as ranking.
Google AI Overviews, ChatGPT and Perplexity all answer a health question by drawing on sources and summarising them. The patient often never reaches a results page at all. Instead of ten blue links they get a short answer with a handful of names in it.
This is not a replacement for local search, and it is not yet the largest channel for most clinics. But it behaves differently enough to be worth preparing for, and the preparation overlaps heavily with work that improves ordinary search anyway, which is why it is included rather than sold separately.
In the audits I run, the clinics that are invisible to AI answers usually share the same causes. There is no structured data on the site, so nothing can be parsed reliably. The FAQ, if there is one, answers operational questions about parking and cancellations rather than the clinical questions patients actually ask. And the practitioner is not named in a way that connects a real, registered person to the advice on the page.
That last point matters more in health than in most sectors. Content attributed to a named, qualified practitioner is a direct expertise signal, and it is one of the criteria used to assess the quality of health content. A clinic with a strong review base and no structured credentials can still be passed over in favour of one that is easier to verify.
Miss all four and a clinic with a strong review base can still be passed over for one that is simply easier to verify.
Anything on your site can be quoted back to a patient by a system you do not control, out of the context you wrote it in. That raises the standard for accuracy rather than lowering it. A loosely worded outcome claim buried in a blog post is a different proposition once it is being read aloud as an answer.
So the content written for this is written to the same standard as everything else here. No guaranteed outcomes, no superlatives that cannot be evidenced, and clinical language that would survive being read by a colleague.
For most UK clinics it is a small share of enquiries today and growing. It is worth preparing for rather than betting on, which is why it sits inside the Growth retainer alongside search work rather than being sold as a separate product.
No, and nobody can. These systems choose sources by their own criteria and change frequently. What can be done is to remove the reasons a clinic gets passed over, which are usually missing structured data, thin condition content, and credentials that are not machine readable.
It overlaps substantially. Schema, condition depth, named expertise and consistent listings help both. The differences are in format, particularly writing answers in the conversational shape a patient would use rather than the keyword shape an agency would.
By checking, which is part of the monthly reporting. Asking the assistants the questions your patients ask and recording whether you appear is crude, but it is honest, and it is more reliable than any tool currently claiming to measure this.