This question turned up at position 31 of 49 in our “people also ask tool” harvest on 27 September 2026, right next to “What is AI SEO called now?”. That’s why it lives in our branch on People Also Ask vs AI Overviews. Google’s own answer, quoting Marcel Digital’s blog, said you “must optimize for Generative Engine Optimization (GEO)”. OpenAI’s documentation is narrower and more modest. Here’s what it actually says.
How ChatGPT search finds pages
OpenAI’s publisher FAQ opens with the good news: “Any public website can appear in ChatGPT search.” ChatGPT may search the web on its own when a question would benefit from current information, or when you pick Search yourself. According to OpenAI’s help article on web search, when it works with other search providers it “typically rewrites your query into one or more targeted queries”. OpenAI’s example: a question about drugs that target CCR8 becomes the search “CCR8 immunotherapy drug development 2025”, followed by narrower searches once the first results come back. The article points readers to Microsoft’s and Shopify’s privacy policies for how those providers handle queries.
OpenAI’s launch post for ChatGPT search (31 October 2024) says it draws on “third-party search providers, as well as content provided directly by our partners”, meaning the news and data providers it has deals with. On top of that, OpenAI runs its own crawler, OAI-SearchBot, which “is used to surface websites in search results in ChatGPT’s search features.” Answers that use search can carry citations and a Sources button listing the pages behind them.
OpenAI’s three bots, and the one that matters
| Bot | What it does | If you block it in robots.txt |
|---|---|---|
| OAI-SearchBot | Finds and surfaces sites for ChatGPT search | Your pages “will not be shown in ChatGPT search answers, though can still appear as navigational links.” |
| GPTBot | Crawls content that “may be used in training” OpenAI’s models | Signals your content shouldn’t be used for training. It does not remove you from search. |
| ChatGPT-User | Visits a page when a user’s request calls for it | “robots.txt rules may not apply”, and OpenAI says it is “not used to determine whether content may appear in Search.” |
The settings are independent. OpenAI’s crawler documentation spells out the exact case: a site can allow OAI-SearchBot to appear in search results while disallowing GPTBot to opt out of training. If you ever blocked “AI bots” as a group, check that OAI-SearchBot didn’t go down with GPTBot.
Going the other way? Blocking OAI-SearchBot doesn’t fully hide a page. The crawler table above already says opted-out sites can appear as navigational links, and OpenAI’s publisher FAQ adds that if it learns a disallowed URL from a third-party search provider or other pages, it may show just the link and page title in ChatGPT Atlas, its browser. To prevent that, it points to the noindex meta tag, with a catch: its crawler has to be allowed to fetch the page to read that tag.
Practical steps
- Check robots.txt. Make sure nothing disallows OAI-SearchBot, including a blanket rule for all user agents. OpenAI says it can take about 24 hours after a robots.txt change for its systems to adjust.
- Check your firewall and CDN. OpenAI’s help article says to confirm your host or content delivery network “allows traffic from OpenAI’s published searchbot IP addresses” (listed at
openai.com/searchbot.json). Bot protection that challenges unfamiliar crawlers can block it without anyone noticing. - Stay indexed in regular search engines. ChatGPT sends rewritten queries to partner search providers, and OpenAI doesn’t publish which provider handles which query. Being well indexed in Google and Bing is the sensible hedge.
- Answer the narrow question, not just the broad one. Rewritten queries are specific. A page with a clear heading for one question and a direct answer in its first sentence or two is easy to match and easy to cite. People Also Ask is a good place to find those follow-up questions, and our guide to turning People Also Ask into a content plan shows how to organize them without spinning up a thin page for every variation.
- Give it a reason to pick you. Original data and first-hand tests are harder to find anywhere else. OpenAI doesn’t publish its ranking factors, so this part is our inference, but it lines up with Google’s advice to write “non-commodity content”.
- Measure it. ChatGPT adds
utm_source=chatgpt.comto referral links from its search results, so those visits show up in Google Analytics or any similar tool.
What nobody can guarantee
OpenAI is blunt about this. Its help article says “ChatGPT ranks search results using multiple factors intended to help users find relevant, reliable information. Placement is not guaranteed.” It doesn’t list those factors. Bing’s Webmaster Guidelines say the same about AI answers in general: “GEO does not guarantee grounding or citations in AI experiences.”
Answers that don’t use search are even further out of your hands. When ChatGPT answers from what its model already learned, there’s no documented way to steer that. Allowing GPTBot means your content “may be used in training”, and that’s where OpenAI’s description stops. Nothing in its documentation says the model will mention you afterwards.
Accuracy isn’t promised either. OpenAI warns: “Search results and citations can be incomplete, outdated, or incorrect.” Ask ChatGPT about your brand or topic now and then, open the pages it cites, and fix what you control. If one of your own cited pages is out of date, update that one first.
So treat any offer of guaranteed ChatGPT placement with suspicion. It promises more than OpenAI does. The groundwork is the same one we describe on what SEO is called now, and it’s why the answer to whether SEO is dead is still no.