Why Your Practice Isn't Showing Up in AI Search Results

When a practice asks why it does not show up in ChatGPT, Perplexity, or Google's AI answers, the expected answer is some missing technical switch. The honest answer is almost always different. AI systems decide who to mention by reading how the rest of the web already talks about you, and for most independent practices, the web is quiet. The work to change that is foundational, not a quick toggle. It also matters more than it used to, because even ranking is no longer the prize it was. Pew Research found that 18% of Google searches already produce an AI summary, and that when one appears, users click a result just 8% of the time versus 15% when none appears (Pew Research Center, 2025). This page is a diagnosis. It walks through the five real reasons a practice stays invisible, and what actually moves it.

The Web Barely Mentions You, and That Is the Real Signal

The instinct, when a practice is missing from AI answers, is to look at the website. Add a schema tag, tighten the keywords, polish the homepage. That instinct points at the wrong layer. The strongest predictor of whether an AI system mentions a brand is not what sits on your own site. It is how often the rest of the web mentions you. Ahrefs, studying 75,000 brands, found that branded web mentions (0.664) correlate much more strongly than backlinks (0.218) with AI Overview brand visibility (Ahrefs, May 2025). Backlinks, the old currency of SEO, barely move it by comparison. Plain mentions of your name, across articles, directories, reviews, forums, and local press, move it far more.

This reframes the whole problem. A practice can have a beautiful site and still be invisible, because being invisible is mostly about the absence of a wider footprint, not a flaw in the site itself. One practice owner we mined in our owner-voice work described her old site as one that 'sat there untouched for years' (Cakesmash owner-voice mining). A site nobody updates, on a web that never references the practice, gives an answer engine nothing to grab onto. The first honest diagnosis is usually this one: you are not being penalized, you are simply not being talked about.

You May Be One of the Invisible Quarter

It is tempting to assume that being missing from one answer is a fluke, that the practice shows up somewhere. Often it does not show up anywhere. In the same Ahrefs study, 26% of brands had zero mentions in AI Overviews, and brands in the bottom 50% of web mentions were described as essentially invisible to AI systems (Ahrefs, May 2025). Roughly a quarter are not under-cited. They are entirely absent. And the dividing line is that web-mention footprint again: fall into the bottom half of how often the web references you, and the systems have nothing to draw from.

An independent practice can land in that bottom half without ever realizing it. The clinical work can be excellent and the patients happy, while the practice's name appears in almost nothing the web has indexed. Another owner from our mining captured the trap from the other side, describing the local field as one where the websites of nearby practices 'pretty much all looked like the same' and she 'did not want my website to look like that, and so I froze' (Cakesmash owner-voice mining). A frozen, undifferentiated presence is exactly the kind that generates no mentions, which is exactly what keeps a practice in the invisible group. The diagnosis here is uncomfortable but useful: not 'we rank lower than we'd like,' but 'we may not exist to these systems at all.'

A Handful of Big Names Hog the Citations

The second uncomfortable fact is concentration. AI answers do not spread their citations evenly. Ahrefs reported that the top 50 brands appearing in AI Overviews account for 28.9% of all citations (Ahrefs, 2025). A tiny roster of the most-mentioned names takes a hugely outsized share of the answers. For broad questions like a general procedure or a category term, those names are usually large institutions and aggregators with web footprints an independent practice cannot match.

This is not a reason to give up. It is a reason to aim correctly. Trying to out-cite a national institution on a generic, high-volume term is the wrong fight. The winnable ground is narrower and more specific: the exact procedure questions and the local provider-selection questions your future patients actually type, where the giant names are thinner and a well-mentioned local practice can surface. The realistic target is not to dominate the broad term. It is to own the specific and the local, where the concentration that buries you on general queries works in your favor instead.

You Are Optimizing Things That Don't Move the Needle

Much of what gets sold as 'AI optimization' is effort spent in the wrong place. Practices add structured-data markup, rewrite headers for keywords, and redesign the homepage, then wait for the AI answers to arrive. They mostly do not, because none of those touch the signal that actually matters. The evidence keeps pointing back to the same place: it is the breadth of mentions across the web, not the tidiness of any single page, that predicts whether you get named. Branded web mentions (0.664) outweigh backlinks (0.218) by roughly three to one in correlation with AI Overview visibility (Ahrefs, May 2025), and the on-page tweaks most practices obsess over barely register.

What does the work is unglamorous and slower. Being mentioned where patients ask questions, in real articles and conversations and local coverage, not just on your own domain. Publishing content that plainly answers the specific questions people ask, in a form an answer engine can lift, rather than a homepage that describes the practice to someone who already decided to book. Our reading of the 834-post content corpus across this niche shows the common pattern is the opposite: a site full of self-description and almost no extractable answers. The misdiagnosis costs real money. A practice can spend a year polishing the surface that AI does not weigh while ignoring the footprint that it does.

Even Ranking Won't Save You If the Click Is Gone

Suppose you fix everything and rank at the top. There is one more honest piece. The click itself is disappearing. Pew Research, watching real browsing data, found that 18% of Google searches produced an AI summary, and that when one appeared users clicked a result on just 8% of those visits, versus 15% when none appeared (Pew Research Center, 2025). When the AI answers the question outright, most people never go anywhere. So a number-one ranking can still send you nothing, because the searcher got what they needed without leaving the results page.

That is precisely why being inside the answer, not merely beneath it, is now the thing that matters. The practices that get named are the ones the web mentions often enough, with answer content clear enough, that the engine pulls them into the summary itself. None of this happens overnight. It is the slow accumulation of a real web presence and genuinely useful answer pages, built over months. There is no shortcut that skips it. The fastest move available is not a trick. It is knowing exactly where your gap sits, so the foundational work goes to the right place instead of the polished but useless one.

The diagnostic frame

The honest summary: most practices are invisible to AI not because of a missing setting, but because the web barely mentions them, and that mention footprint is what these systems read. A quarter of brands get cited zero times, and the bottom half of web-mention volume is essentially invisible (Ahrefs, May 2025). The fix is foundational and takes time. Before spending on any of it, the useful first step is a clear diagnosis of where, specifically, your practice falls short.