Most sites missing from ChatGPT citations are failing at one of six checkpoints: crawler access, Bing indexing, content extractability, domain authority, content freshness, or simply not measuring whether they're already being cited. Unlike Google, ChatGPT Search doesn't run on its own index, it retrieves from Bing, so a citation failure can originate somewhere you wouldn't normally check. The order below runs from the most common false alarm to the most commonly skipped fix.
You're probably not blocked, but check the right bot
Before assuming a technical block is the problem, it's worth knowing how rare that actually is. A July 2026 audit by Inspeccia pulled robots.txt from 1,838 domains and got 1,567 valid files back. Of those, exactly nine (0.6%) fully blocked a crawler that actually determines ChatGPT search citations (Inspeccia). Compare that to the 108 sites (8.2%) blocking GPTBot, OpenAI's training crawler, which has no bearing on search citations at all.
The confusion is understandable: OpenAI runs three separate bots with similar names and very different jobs. GPTBot collects training data. ChatGPT-User fetches a page live when a person asks ChatGPT to look at something specific. OAI-SearchBot is the one that actually determines whether you can be cited in ChatGPT's search results, and per OpenAI's own FAQ, blocking it removes you from those answers, full stop. Of the 108 sites in the Inspeccia audit that blocked GPTBot, 91 of them (84%) never touched OAI-SearchBot at all. They kept their content out of training and left the citation door wide open, usually without meaning to.
So before you go further down this list: open your robots.txt and search specifically for OAI-SearchBot. If it's absent, you're allowed. If it's there with a Disallow: / under it, that's a two-minute fix. Two other places worth a look, since neither shows up in robots.txt: OpenAI's documentation also asks that you allow requests from its published IP ranges, and if you're on Cloudflare, every domain registered since July 2025 defaults to a control setting for AI crawlers that can block OAI-SearchBot at the CDN level regardless of what your robots.txt says.
1. Are you indexed on Bing at all?
This is the eligibility check most sites skip, because it doesn't occur to Google-only teams to look at Bing. ChatGPT Search retrieves from Bing's index, not Google's, a Seer Interactive study found 87% of ChatGPT citations overlap with Bing's top-10 organic results (cited in our AI search optimization guide). If you've never submitted a sitemap to Bing Webmaster Tools, or if Bing's crawl errors for your domain have gone unchecked, you may simply be invisible to the index ChatGPT is drawing from, independent of how well you rank on Google.
2. Is your content structured for extraction?
The same extraction problem that affects Google AI Overviews applies here: ChatGPT quotes specific passages, not full pages, so content that buries its answer under throat-clearing is much harder to lift into a response. Lead each section with a direct two-to-three sentence answer, use question-or-answer-shaped headings, and keep the structure scannable. This is also where sourcing matters, cite primary sources explicitly and quote them directly, since both Bing's ranking signals and ChatGPT's own preference for verifiable claims reward it.
3. Does your domain carry enough authority?
Ahrefs analyzed the top 1,000 pages ChatGPT cited in a given month and found that among the cited pages that also rank organically, 65.3% sit at Domain Rating 81 or higher, with a median DR of 90 (Ahrefs). Interestingly, the specific page's own authority barely mattered: most had very few backlinks of their own (median URL Rating of just 6). The takeaway: ChatGPT leans heavily on whether your domain overall is established, not whether the individual page has been heavily linked to. A strong, well-linked homepage and site-wide authority carries new or thin pages further than you'd expect from their own link profile alone.
4. Is your content stale?
Of the ChatGPT-cited pages in the same Ahrefs study where a freshness date could be detected, 76.4% had been updated within the last 30 days. That's from a smaller sample: 564 of the 1,000 pages had a detectable date, and Wikipedia (which updates constantly) made up over half of that subset, so treat it as a strong directional signal rather than an exact rate you should expect to hit. Still, the pattern held even after Wikipedia was excluded from the analysis, with roughly 82% of the remaining pages still updated within the past year. If your highest-value pages haven't been touched in over a year, that's worth addressing before anything else on this list.
5. Are you actually checking whether you're being cited?
Unlike Google, ChatGPT gives you a real signal here: it appends utm_source=chatgpt.com to links in its responses. The catch is that most teams never set up the filter. Add a segment in Google Analytics or your analytics platform for that UTM parameter and you'll get an accurate count of ChatGPT-referred sessions, which you can break down by landing page to see which content is actually driving citations.
That said, a UTM filter only tells you about citations someone clicked. It won't tell you about the (often much larger) number of times ChatGPT named or quoted your site without a click, or how you're doing in Google AI Overviews and Claude, which don't offer anything comparable to a UTM parameter at all. That gap (knowing whether you're cited, mentioned, or absent across all three platforms, not just the one with decent tracking) is what tools like Citation Hawk are built to close.
Putting it together
Rule out the crawler confusion first, it's the fastest check and the least likely culprit despite being the most talked-about. Then confirm you're actually indexed on Bing, check that your content is structured for extraction, weigh in on domain authority versus page-level authority, look at how stale your top pages are, and finally, set up the one piece of tracking ChatGPT already gives you for free.
For the full technical checklist (covering ChatGPT Search, Google AI Overviews, and Claude in one place), see our guide to getting your website into AI search results. If Claude is also missing you, the checkpoints are similar but not identical: see why Claude might not be citing your site.
Tom Eastwood is the founder of Citation Hawk, a tool that monitors AI citation rates across ChatGPT, Claude, and Google AI Overviews.