Can a brand-new website get cited by AI search engines?

A new site can be cited. The real question is which engine, and that depends on two gates the site has to pass with each one.

The first gate is access: can the engine's crawler fetch the page? The second is the index: does that engine, or the search index it draws on, hold the page? Only after both does retrieval decide, query by query, whether your page enters the answer. Then the model picks what to cite from what it retrieved. You control the first two gates. You don't control the last two.

Google states this plainly for its own surfaces. To appear 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," and Google says there are no additional technical requirements. It lists no age threshold. More on that in whether domain age matters for AI search.

One gate a new site can't pass yet is model memory. Answers given without live search come from training data frozen at a cutoff. OpenAI's model docs, checked in August 2026, list a February 16, 2026 cutoff for the GPT-5.6 family and August 31, 2025 for the Instant path ChatGPT uses. A site launched after those dates only reaches answers that search the web.

OtterlyAI ran the clearest public test of new domains in 2026. It launched two brand-new sites about capybaras, each with roughly 1,000 AI-generated posts, and tracked them for about four months across seven AI engines.

• Google removed both. Otterly reports one site lasted about 14 days in Google's index and the other about 90, with zero manual actions in Search Console.
• Bing kept both. Otterly reports combined Bing impressions grew every month, to 119,935 in July.
• Copilot kept citing. Of 3,934 AI citations Otterly tracked for one site, 2,718 (69%) came from Microsoft Copilot, which grounds on Bing. That site's citations rose from 100 in April to 1,666 in the first 27 days of July, while it was gone from Google.
• Google's AI surfaces went quiet. AI Overviews, AI Mode and Gemini almost never cited the sites.
• ChatGPT cited effectively nothing, even with its crawlers visiting and the sites ranking in Bing.

The lesson: an engine can't cite what its index no longer holds, and being indexed somewhere is necessary but not enough. Read it with the caveats. It was one site per arm, a low-stakes informational topic, and a vendor that sells AI tracking ran it.

Run these before you worry about content. Each one removes a block, and a single failure can hide the site from an engine entirely.

1. Allow the search crawlers in robots.txt, per bot. OAI-SearchBot controls ChatGPT search. GPTBot controls training only, and OpenAI says "each setting is independent of the others." Check PerplexityBot, Claude-SearchBot, Googlebot, bingbot and Meta-WebIndexer too.
2. Confirm the firewall or CDN lets them through. OpenAI's help centre says eligibility needs the crawler allowed and the host or CDN allowing its published IP addresses. From September 15, 2026, Cloudflare blocks Training and Agent crawlers by default on ad-displaying pages for new domains, while Search stays allowed. A 403 is a silent failure. The AI Crawler Checker requests a page as each AI crawler and shows which ones get turned away.
3. Put key facts in the HTML before JavaScript runs. Vercel found ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. Googlebot renders; most AI crawlers don't.
4. Remove accidental suppression. Templates often ship noindex, nosnippet, max-snippet:0 or data-nosnippet. Each keeps content out of AI answers.
5. Verify the index engine by engine. Check Google Search Console and Bing Webmaster Tools separately. Otterly's sites were healthy in one and gone from the other.
6. State specific facts plainly and consistently. A price or a one-line definition on a fetchable page gives an engine a sentence it can quote. Contradictions across your own pages give it the wrong one, and the free brand fact consistency check finds them.

For Google surfaces, see getting cited in AI Overviews and AI Mode. For ChatGPT, see what you control in ChatGPT.

Volume looks like a fast way to give a new domain something to cite. The Otterly test shows the cost. Google's spam policy treats it as scaled content abuse, and its first example is "using generative AI tools or other similar tools to generate many pages without adding value for users." Neither site got a warning in Search Console.

The tool isn't the problem. Google says "appropriate use of AI or automation is not against our guidelines." Volume without value is.

Fake expertise didn't help either. The site with a fictional wildlife biologist in its author box was deindexed first. One site can't prove author boxes hurt, but it's no evidence they protect. See whether author bios help AI citations.

Copilot citations don't make up for losing Google. Publish fewer pages that answer real buyer questions.

A new domain has few pages and no history in any index. The evidence shows a strong association between brand prominence and citation, with no lever to manufacture it. When engines cite other sites for your category, being named on those sites is the route into the answer. Start with how third-party mentions reach AI answers, and if you have no coverage yet, read how a startup with no press gets mentioned.

Then measure each engine on its own. One engine cited Otterly's sites thousands of times while four barely did, and a blended score would have hidden that. The free checker on this site shows whether engines name you for up to three buyer questions. It's a snapshot, so rerun it after each fix.