Can smaller, lower-traffic websites get cited by AI search engines?

Can smaller websites get cited by AI search engines?

Yes, and often. Many cited sites receive only one or two citations. Zhang and colleagues (opens in a new tab) compared six AI engines with Google and Bing on 55,936 queries in July and August 2025. Except for ChatGPT, the AI engines tended to cite less popular domains than traditional search.

The gap was largest for two engines. Copilot and Gemini cited domains whose average traffic rank was over 22,000 places lower than Google’s and Bing’s. The authors also found that 37% of all cited domains appeared only in AI answers, never in traditional results.

The traffic ranking used here is a research list of website popularity. It measures whole domains, not individual pages.

Do AI engines favor big, popular websites?

The evidence is mixed and depends on the engine. A global study by Aral and colleagues (opens in a new tab) ran 24,000 search queries in 243 countries in 2024 and 2025. It found Google’s AI Overviews linked to the long tail of the web significantly less than traditional search, and to the top 1,000 sites more.

Grossman and colleagues (opens in a new tab) found the opposite pattern for the first source shown. In traditional Google results, the top source came from one of the 1,000 most popular sites for 52.7% of queries. For AI Overviews the figure was 40.0%, and for Gemini 32.6%.

These studies used different query sets, popularity lists and years. The honest summary is that no engine reliably shuts out smaller sites, and none reliably favors them. For the kinds of sites engines lean on, see which websites AI search relies on most.

How spread out are AI citations?

Very spread out, with a short list of repeat winners. In the Xu study of Google AI Overviews in spring 2026, 56.2% of cited sites were cited exactly once over 40 days. On Google’s first page of results the share was 52.5%.

The top of the list is less dominant in AI Overviews than on Google’s results page. The five most-cited sites took 20.0% of AI Overview citations, against 47.0% of first-page links.

News is the exception. Yang (opens in a new tab) found that for OpenAI models in 2025, the 20 most-cited news outlets took 67.3% of citations in the news sample. A small publisher competing on news faces a far more concentrated field.

Is being cited the same as being recommended?

No. A small site can be cited as evidence while its brand is left out of the answer. Kumar (opens in a new tab), whose company sells AI visibility tracking, measured more than 100 brands between March and May 2026. Household names appeared in 73% of relevant unbranded answers on the first run, mid-market brands in 44% and niche brands in 11%.

Other studies show the same pull toward big names. Chen and colleagues (opens in a new tab) asked ChatGPT and Perplexity 50 unbranded cola questions in 2025. Niche brands were 12.3% of ChatGPT’s brand mentions and 5.8% of Perplexity’s.

New companies face the steepest climb. Sharma (opens in a new tab) tested 112 Product Hunt startups. A version of ChatGPT without web search recognized them 99.4% of the time when asked by name. It surfaced them in only 3.32% of discovery questions such as “What are the best AI tools launched this year?”.

What helps a smaller site get cited?

Depth, relevance and visible proof, more than size. Zhu and Chang (opens in a new tab) compared 14 Tokyo hotel websites against Gemini’s citations. One small independent hotel with no schema markup won direct citations through a 33-question FAQ and a detailed sightseeing guide. A polished boutique with shallow content did not. The audit is small and exploratory.

Our own data shows small creators getting in. In our YouTube study, videos cited from outside page one had a median of 2,776 views, against 34,100 for cited videos Google also showed. Within the same search, a bigger channel lowered the chance of citation.

For local businesses, reviews matter. In our ChatGPT local picks study, businesses with more reviews than the local median were 19.5 points more likely to be listed.

Small advantages can matter in tests too. In a simulated skincare study by Chu and Hou (opens in a new tab) using three AI models, well-known brands won 100% of recommendations when products were otherwise identical. That lead disappeared when a rival had a rating advantage of less than 0.1 stars.

Can optimization close the gap with bigger sites?

Possibly, in simulations. The original GEO study by Aggarwal and colleagues (opens in a new tab) tested content rewrites on a simulated engine. Adding citations to sources raised visibility by 115.1% for websites ranked fifth in search, while the top-ranked website lost 30.3% on average.

Live evidence is weaker. In Sharma’s startup study, scores for AI-oriented page optimization showed no link to discovery. For Perplexity, traditional signals such as links from other sites did.

What should you do about it?

Compete where size matters least: specific questions, deep answers and proof from others.

1. Target narrow, specific questions where large sites have thin coverage.
2. Answer them in depth on your own pages, with concrete facts, FAQs and guides.
3. Publish short, useful videos. Google’s AI cites small channels.
4. If you serve a local market, build review volume on your Google profile.
5. Earn links and mentions from other sites, which help discovery on engines that search the web.
6. Measure brand mentions, not only citations, because the two diverge for smaller brands.

Our generative engine optimization service works on this kind of focused visibility.

What does the research not tell us yet?

No study we reviewed followed small sites over time to see what moved them from cited to recommended.

• Studies disagree on whether AI engines favor popular sites, and they use different popularity lists.
• The brand-size figures come from a vendor dataset and small tests, not independent large samples.
• The optimization gains come from simulated engines, not live ones.
• Several samples are narrow: Product Hunt startups, Tokyo hotels or cola brands.
• In our brand entity study, 21 of 110 options all four assistants named had no Wikipedia article, and 11 were local businesses. Why such businesses break through is still unexplained.

Frequently asked questions

Do AI search engines only cite big websites?

No. In one study of AI Overviews, 56.2% of cited sites were cited only once in 40 days, which points to a long tail of smaller sources.

Does ChatGPT prefer popular websites?

More than most AI engines. In a 2025 study of six engines, ChatGPT was the only one that did not cite less popular domains than Google and Bing.

Can a new startup show up in ChatGPT answers?

Rarely at first. One study found a version of ChatGPT without web search surfaced Product Hunt startups in 3.32% of discovery questions.

Is YouTube a way for small brands to get into AI Overviews?

It can be. In our study, cited videos from outside page one had a median of 2,776 views, far fewer than the videos Google ranked.

Sources

• Zhang et al. (2025), Source Coverage and Citation Bias in LLM-based vs. Traditional Search Engines (opens in a new tab), arXiv:2512.09483.
• Aral et al. (2026), The Rise of AI Search: Implications for Information Markets and Human Judgement at Scale (opens in a new tab), arXiv:2602.13415.
• Grossman et al. (2026), How Generative AI Disrupts Search: An Empirical Study of Google Search, Gemini, and AI Overviews (opens in a new tab), arXiv:2604.27790.
• Xu et al. (2026), Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact (opens in a new tab), arXiv:2605.14021.
• Yang (2025), News Source Citing Patterns in AI Search Systems (opens in a new tab), arXiv:2507.05301.
• Kumar (2026), Generative Engine Optimization at Scale: Measuring Brand Visibility Across AI Search Engines (opens in a new tab), arXiv:2606.20065.
• Chen et al. (2025), Generative Engine Optimization: How to Dominate AI Search (opens in a new tab), arXiv:2509.08919.
• Sharma (2026), The Discovery Gap: How Product Hunt Startups Vanish in LLM Organic Discovery Queries (opens in a new tab), arXiv:2601.00912.
• Zhu and Chang (2026), The End of Rented Discovery: How AI Search Redistributes Power Between Hotels and Intermediaries (opens in a new tab), arXiv:2603.20062.
• Chu and Hou (2026), Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems (opens in a new tab), arXiv:2606.17443.
• Aggarwal et al. (2024), GEO: Generative Engine Optimization (opens in a new tab), arXiv:2311.09735.
• Underneath (2026), The YouTube videos Google’s AI cites are small
• Underneath (2026), Which Google Maps businesses does ChatGPT recommend?
• Underneath (2026), Do Wikipedia and schema make AI assistants recommend a brand?