If ChatGPT or Perplexity never cite your site, the cause usually sits in one of four checkable buckets — starting with whether the 16 named AI crawlers (GPTBot, ClaudeBot, PerplexityBot and 13 more) can even reach your pages. Each bucket below is something you can check yourself in minutes, without ranking which one is more common — there's no reliable rate for that.
Four checkable causes, in this order
These four causes are not ranked by frequency. Walk through them in order; each one rules itself in or out before you move to the next.
| Cause | What it is | Check |
|---|---|---|
| 1. Crawler access | AI crawlers are blocked or can't fetch pages | robots.txt + real fetch test |
| 2. Blank to AI | Page content only renders via JavaScript | Raw HTML check for known text |
| 3. Nothing quotable | No category, area or hard numbers to lift | Read your own page as an answer |
| 4. One source layer only | Site-only effort, no third-party or community presence | Search your brand + topic off-site |
Cause 1: can the crawlers even reach your pages?
This is a yes/no gate, not a ranking factor — passing it earns nothing by itself, but failing it ends the conversation before it starts. There are 16 named AI crawlers (GPTBot, ClaudeBot, PerplexityBot and 13 more) that need to fetch your pages; if robots.txt blocks them or your server returns an error to their user-agent, nothing downstream matters. Check it in minutes: pull up your robots.txt and look for disallow rules naming any of these crawlers, then run a real fetch against a few key pages. The full checklist is in the AI crawler access checklist. What to do: remove any blocking rules and confirm with a second fetch.
Cause 2: does the page read blank to AI?
Many AI crawlers do not execute JavaScript, so a script-built page reads as an empty shell even though it looks complete in a browser. This only applies to script-rendered sites — it is not the default explanation for every site, and it never doubles as the most common cause. The check takes about 30 seconds: fetch the raw HTML of the page (not the rendered view) and search for a sentence, phone number or product name you know is on the page. If it's not in the raw HTML, the crawler never saw it either. Details and fix options are in why AI sees a blank page. What to do: server-render or statically generate the content that matters for citation.
Cause 3: is there anything on the page worth quoting?
Pages that are crawlable and fully readable can still go uncited if they give the engine nothing to lift — no category, no service area, no hard numbers, just general claims. There's no statistic on how often this specific cause is the reason; it's a question you answer by reading your own page the way an engine would. Check: open your top page and ask whether a sentence could be copied verbatim into an answer. What to do: add concrete, checkable specifics — what you do, where, and a number that matters to the reader.
Cause 4: are you covering only one of three source layers?
AI answers draw on three source layers — your own site, third-party authority sources, and community discussion — and a site-only plan covers one of them. See where AI answers actually come from for how the three layers work together. This explains why a well-built site can still be absent on open questions: the engine may be pulling from a review site or a forum thread instead of your homepage, regardless of how clean your own pages are. The three layers are described here without ranking one above the others — each plays a different role depending on the question.
Why a perfect technical score does not fix it
In one controlled comparison, a site scoring 92/100 on technical readiness got zero AI citations, while a technically messy rival was cited in 3 of 3 runs. That's a single comparison, not a percentage or a rule — but it's a useful check against treating a technical checklist as the finish line. Technical readiness (crawler access, clean HTML) is a prerequisite, not a guarantee of citation. Full writeup: technical score is not citations.
What to expect after fixing any of the above
Fixing one of the four causes does not produce an instant change, and testing again after a day or two will often look like nothing happened. Crawler behavior changes within days to weeks of a fix. Citations in answers usually take 2–8 weeks to show up once the underlying issue is resolved. A brand-new domain needs about 3 months of observation on top of that, because trust signals for a new domain build up more slowly. These are observation windows, not promises — more in how long until AI mentions you.
AI search engines ignore my site completely — what am I doing wrong?
If it feels like ChatGPT or Perplexity ignore your site completely, the diagnosis is the same four checks above, worked in the same order — there's no separate explanation for "completely ignored" versus "occasionally missing." Start with crawler access, then the blank-page check, then whether your pages contain anything quotable, then whether your presence sits only on your own site. Nothing here guesses what's wrong with your specific site beyond those four checks; running them is how you find out.
First step: the free 24-item AI-visibility audit
A 24-item AI-visibility audit (free) covers checks (1) and (2) above — crawler access and raw-HTML readability — without needing technical skills to run it yourself. It checks access and readability; it does not measure mentions or citations, since those depend on the other two causes and on time passing after a fix. It's a starting point for ruling causes in or out, not a report on whether you're already being cited.
FAQ
Why does ChatGPT never mention my business even though my website works fine?
"Works fine" in a browser doesn't rule out the four causes above — a site can render perfectly for a human and still be blocked for crawlers, blank in raw HTML, short on quotable specifics, or simply absent from the third-party and community layers that AI answers also draw on. Check crawler access and raw HTML first; those two take minutes.
How long should I wait after fixing something before I check again?
Crawler behavior typically changes within days to weeks, but citations in actual answers usually take 2–8 weeks, and a brand-new domain needs about 3 months of observation on top of that — see the full timeline breakdown. Testing again after one day is too early to conclude anything.
My technical SEO score is high — why does Perplexity still not cite me?
A high technical score is not the same as being quotable: in one controlled comparison a site at 92/100 technical readiness got zero citations while a messier rival was cited in 3 of 3 runs, detailed in this comparison. Technical cleanliness is a gate, not a guarantee.