Is Your Site Invisible to AI Search? How to Tell and Fix It

I sat down to check a client's AI visibility a while back, expecting to find some exotic GEO problem. Brand new category, sharp positioning, the kind of company that should turn up when you ask an assistant who does this work. It didn't. Not in ChatGPT, not in Perplexity, not in Google's AI answers. The founder assumed the AI engines just hadn't caught up to them yet.

The actual cause was boring. A handful of their most important pages weren't indexed in Google at all. No index, no snippet, no AI citation. The mystery dissolved into a checklist.

That pattern is common enough that I want to give you the checklist directly. Most people treat AI invisibility as a black box, something happening inside a model nobody can inspect. Sometimes that's true. More often, the reason is visible in tools you already have, and the diagnosis takes an afternoon, not a research project.

One distinction up front, because it saves a lot of wasted effort. Being absent from AI answers and being described wrongly in them are two different problems with two different fixes. This piece is about absence: the engines don't surface you at all. If AI is surfacing your business but getting the facts wrong, that's entity confusion in AI search, and the diagnosis runs a different way. Being wrong is one problem. Being absent is another. Know which one you have before you start fixing.

How do you tell if you're invisible to AI search?

Start with the one thing that is documented rather than guessed at. For Google's AI features, the entry requirement is plain. In its AI Features and Your Website documentation (updated December 2025), Google states that "to be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet." There is no separate AI index. There is no special markup that gets you in. If the page is indexed and snippet-eligible, it's in the running. If it isn't, it can't appear, full stop. To check this, open Google Search Console's URL Inspection tool and inspect your most important pages one by one. You're looking for "URL is on Google," no accidental noindex, and no nosnippet or a tiny max-snippet limit quietly suppressing the snippet. Google names those snippet controls explicitly as the levers that keep a page out of AI features.

People apply snippet restrictions by accident more often than you'd think, usually a leftover from a staging environment or a plugin default nobody revisited.

This is the part most "how to appear in AI search" advice skips, because it isn't exciting. It's also the part that's actually documented by the company running the largest AI search surface. Google's own guide to optimizing for generative AI features (updated June 15, 2026) is blunt about it: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Same guide, important caveat worth internalizing: "just because a page meets all requirements, best practices, and complies with the policies, doesn't mean that Google will crawl, index, or serve its content." Indexing makes you eligible. It does not promise a citation. Treat it as the floor, not the finish line.

What is the AI invisibility checklist?

Here's the full diagnostic in order. Run it top to bottom. Each row tells you how to check and, where the answer is "you have a real problem," where to go to fix it. The first two rows are the documented gate. The rest narrow down why an indexed page still isn't getting cited.

A note on the manual query test, because it trips people up. Results vary session to session. Ask the same question twice and you can get two different answers. Run your checks across several days and several sessions before you conclude anything, and ask in a few shapes: "What can you tell me about [your brand]?" and "Who are the leading [your service] in [your city]?" One bad session isn't a verdict.

Reading your robots.txt for AI crawlers

The crawler row deserves its own breakdown, because the rules differ by engine and people get this backwards constantly. The most damaging mistake I see is a blanket User-agent: * disallow that someone added for one reason and forgot, which can quietly wall off the crawlers that feed AI answers.

What is the difference between being absent from AI search and being described wrongly?

This is worth slowing down on, because the two problems feel identical from the outside and need opposite fixes.

Absent means you ask an assistant a question your business should answer, and you're simply not there. No mention, no link, nothing. The diagnosis runs through the checklist above: indexing, crawler access, snippet eligibility, presence across the web.

Wrong means you are there, but the facts are off. The model names you but attaches the wrong service, an old address, a competitor's detail, or merges you with a similarly named company. That's a recognition problem, and it has its own diagnosis. I've written about it separately in the piece on entity confusion in AI search. If that's your symptom, start there instead. Chasing indexing fixes won't help a brand that's already indexed and just being misread.

Quick gut check. Open ChatGPT or Perplexity and ask about your business. No answer at all, or "I don't have information about that"? You're likely dealing with absence. A confident answer that's partly wrong? That's the entity problem. Different door.

Why does AI invisibility matter even if you rank well in search?

You might be reading this thinking your SEO is fine, so AI visibility is a bonus you'll get to later. I'd push back on that, and not because GEO is some separate discipline you have to bolt on. For how I work, search and AI visibility are the same job. Google itself frames it that way in the June 2026 guide: optimizing for generative AI search is optimizing for search, and thus still SEO. SEO is AEO is GEO. They're one practice.

The reason it's urgent is what's happening to clicks. A SparkToro study using Similarweb clickstream panel data, covering January through April 2026, found that 68.01% of U.S. Google searches ended without a click to any website, up from 60.45% in 2024. For every 1,000 U.S. Google searches, only 276 clicks now reach the open web. AI Overviews are the main driver. The methodology excludes Google's mobile app searches, where the number is likely higher still.

Put those together. More searches resolve inside the answer, and the answer increasingly is an AI summary. Semrush's AI Overviews study, published December 15, 2025 and based on more than 10 million keywords analyzed from January through November 2025, found AI Overviews appeared in 6.49% of searches in January 2025, peaked at 24.61% in July, and settled to 15.69% by November. Prevalence figures swing a lot across trackers depending on keyword set and method, so I'm date-stamping that one deliberately. The direction is the point: if you're not eligible to be cited in those answers, you're invisible in a growing share of the results page, even on queries you "rank" for.

One more, with a caveat attached. A vendor study from Loamly, published February 27, 2026, analyzed 2,089 brands and reported that 77% were completely absent from AI platform responses. I'd treat that as one data point from a small study, not a settled industry figure. But it squares with what I see: absence is the default, not the exception. Most businesses haven't cleared the basic gate yet.

What should you do if your site is invisible to AI search?

Once the diagnosis points to a cause, the fixes are specific. I won't re-teach each one here, because I've already written them up in depth. The job of this page is to get you to the right fix, not to repeat it.

If pages aren't indexed or snippet-eligible: this is foundational technical SEO. Clear crawl errors, remove stray noindex and nosnippet directives, confirm the pages meet Search Essentials, and make sure key content is in server-rendered HTML rather than locked behind JavaScript. If pages are leaving the index entirely, that's its own diagnosis worth reading: why Google deindexes pages. And if this is a newly built or relaunched site that never showed up properly in regular search to begin with, that's a different problem with its own patterns, which I've walked through in why new websites don't show up in search.

If your entity signals are weak: this is about making it obvious who you are, consistently, in a form engines can resolve. Valid Organization schema with sameAs is part of it, and the way to add it shouldn't require a quarter of developer time. I've made that case in why adding schema shouldn't be a dev project. A quick word on expectation-setting, since the brief honesty matters here: Google does not say "add Organization schema and you'll appear in AI Overviews." Schema earns rich results and clarifies your entity. The link to AI citation is reasonable inference from how retrieval works, not a Google promise. I'd rather you know that than chase schema as a magic switch.

If you're curious about the newer entity and access standards: two proposals come up constantly in these conversations. One is a machine-readable entity map, which I've broken down in what an entitymap is. The other is the llms.txt file. Worth understanding both, with one honest caveat from Google's June 2026 guide: for Google Search specifically, "you don't need to create new machine readable files, AI text files, markup, or Markdown," and llms.txt is named directly as not needed for Google. Other engines may treat these differently, and the standards are young. Read those two pieces for the full picture before you invest in either.

If you're absent despite a clean technical setup: this is usually the off-site problem, and it's the one most businesses underinvest in. If your brand only exists on your own domain, AI engines have nothing independent to corroborate. They draw confidence from presence across many sources, not from how polished your own site is. That means earned mentions: directories, third-party reviews, guest content, and for B2B especially, your presence on platforms the engines lean on. I've written specifically about how LinkedIn drives AI citations for B2B, which is a concrete place to start.

If you walk the checklist and your pages are indexed, snippet-eligible, crawlable, and you still can't find yourself in AI answers, that's the point where the work gets genuinely strategic. Untangling entity signals and building the off-site presence that earns citations isn't a single fix you check off. It's a discipline, and it's the kind of work a search and AI visibility engagement is built to carry. If you've run the diagnosis and the answer is "everything looks right and I'm still not there," that's a conversation worth having.

And if you want the broader picture of how to get search help, who does this kind of work, and how to choose, the pillar this sits under walks through it: do you need an SEO consultant. AI visibility is one of the high-stakes, high-judgment problems that piece is about.

Invisibility to AI search usually isn't a model problem. It's a chain you can inspect, and most sites break it early.