// Answer
Why aren’t my pages cited by ChatGPT?
Almost always one of five mechanical things: you’re chasing the wrong KPI, the crawler can’t read you, your schema is thin, your entity is ambiguous, or nobody trusted is talking about you yet. Diagnose in that order.
// Step 1
Confirm you’re trying to be cited, not ranked.
Citation and rank are different objectives. Citation is “the engine named me inside the answer.” Rank is “my URL is on a list of links.” ChatGPT mostly does the first and only sometimes does the second — and a brand that’s great at ranking can be terrible at being cited because it relies on link-bait pages, vague positioning, and JavaScript frameworks. Decide which KPI you’re solving for before you start changing things. The fix sets overlap, but they’re not identical.
// Step 2
Check that crawlers can actually read your page.
The most common silent failure. Open your robots.txt and look for explicit disallows against GPTBot, ChatGPT-User, OAI-SearchBot, PerplexityBot, ClaudeBot, and Google-Extended. Some teams blocked these in 2024 and never went back. Then load a representative page in a browser with JavaScript disabled. If your headline, body copy, and key facts disappear when JS is off, the in-browsing crawler that ChatGPT uses to extract spans can’t see them either. Server-side render or use a static build for content pages. Whether AI engines respect robots.txt covers the crawler side in more depth.
// Step 3
Check your schema.
If your high-value pages don’t emit JSON-LD, you’re forcing the engine to guess what they’re about. Use a schema validator and confirm each high-stakes page emits the right type: Organization and WebSite at the root, Article on content pages, FAQPage on anything with question-and-answer structure, Product and Offer on commercial pages, DefinedTerm on glossary entries, BreadcrumbList for navigation. The goal isn’t schema for SEO’s sake. It’s handing the model your facts pre-parsed so it doesn’t have to extract them under pressure.
// Step 4
Check your entity completeness.
Open Wikidata, Crunchbase, LinkedIn, G2, and your About page side by side. Is your legal name the same on all five? Is your category claim the same one sentence in all five places? Are your founders, headquarters, and founding date consistent? Brands that fail this audit usually have the same problem: the engine resolves them as two or three different entities, or as a single entity with conflicting descriptions, and silently demotes them. The fix is consistency, not creativity — same name, same category, same one-line description, same domain across every public surface that describes you.
// Step 5
Check your link and citation graph.
If steps 2–4 are clean and you’re still not being cited, the bottleneck is almost always here. The engines you’re trying to win lean on the same handful of trust signals the rest of the web does: who links to you, who reviews you, who writes about you on G2, Reddit, Hacker News, Wikipedia, and established trade press. If you’re a category newcomer with zero of those, no amount of schema will rescue you in the short term. Earn the citations off-site, then watch them feed back into the on-engine ones. Tracking AI citations is how you tell whether the inputs are converting into outputs. Start free to see your current footprint.
FAQ
Common questions.
How do I know if ChatGPT can even crawl my site?
Check your robots.txt for explicit blocks against GPTBot, ChatGPT-User, and OAI-SearchBot. Then load a representative page with JavaScript disabled — if your main content disappears, ChatGPT’s parsing path can’t see it either, even when browsing is on.
Is being missing in ChatGPT the same as not ranking on Google?
No, though they correlate. You can rank on Google and still not be cited by ChatGPT — the engine extracts and rewrites rather than linking, and it weighs schema, entity clarity, and content structure differently. The fix set overlaps but isn’t identical.
How much schema is enough?
Organization and WebSite at the site level. Article on every content page. FAQPage on pages with Q&A. Product and Offer on commercial pages. BreadcrumbList everywhere. If your highest-stakes pages have none of those, schema is your first fix, not your last.
Why does my brand resolve as the wrong thing?
Usually entity collision — another company with a similar name, or your old name still cached, or inconsistent category claims across listings. The fix is consistency: same legal name, same one-line category, same domain in your Crunchbase, LinkedIn, Wikidata, and Organization schema. Force the engine to disambiguate in your favor.
How long after I fix things should I expect citations to appear?
Browsing-path citations can shift within days. Training-data citations only refresh on new model versions. Most teams see meaningful movement in 2–6 weeks for browsing-tool answers and across a quarter for steady-state visibility.