When ten sources say the same thing about a topic, AI doesn't cite all ten. It cites the one that adds something the other nine don't have. Google's Information Gain patent describes this explicitly. The patent measures the additional value a document provides beyond
In 2024, author credentials carried approximately 8% weight in AI citation decisions according to BrightEdge's tracking. By 2025, that number doubled to 16%. The weight of author signals in AI search is growing faster than almost any other factor being measured. If your content doesn't have
In July 2025, roughly 76% of the pages cited in Google AI Overviews also ranked in the top 10 organic results for the same query. SEOs could treat ranking as a reasonable proxy for AI visibility. If you ranked well, you'd probably get cited. Eight months later, that
If you're optimizing your content for ChatGPT and assuming that covers AI search, the data has bad news. Only 11% of domains are cited by both ChatGPT and Perplexity. And just 12% of URLs cited by ChatGPT, Perplexity, and Copilot rank in Google's top 10. At
SE Ranking studied 129,000 domains and 216,524 pages and produced one of the cleanest findings in GEO research: pages with First Contentful Paint under 0.4 seconds averaged 6.7 ChatGPT citations. Pages with FCP above 1.13 seconds? Cited three times less often. That's not
For roughly a decade, domain authority has been the metric SEO professionals, such as Radiant Elephant, use to predict competitive outcomes. Higher DA means better rankings. Better rankings mean more traffic. The logic was clean, and the metric was measurable. Then AI search showed up and mostly ignored it. Wellows&
Here's a contradiction that confuses a lot of GEO practitioners: Semrush's study of 304,805 URLs found a +22% citation lift associated with structured data, making it the fifth-strongest predictor of AI citation. But Search Atlas's study of 5.5 million responses across
Vercel and MERJ tracked over 500 million GPTBot fetches and found something that should concern every site owner relying on client-side JavaScript: zero evidence of JavaScript execution. None. GPTBot downloads JavaScript files 11.5% of the time, but it never runs them. The same is true for ClaudeBot, PerplexityBot,
Wikipedia accounts for 7.8% of all ChatGPT citations. That single domain represents nearly half (47.9%) of ChatGPT's top-10 most-cited sources. Across virtually every major LLM, Wikipedia sits at #1 or #2 for citation frequency. This matters because of how AI retrieval actually works. Before
YouTube is the #1 most-cited domain in Google AI Overviews. Not Wikipedia. Not Reddit. Not Forbes. YouTube. Ahrefs Brand Radar tracks citation sources across AI platforms, and their data shows YouTube's citation share in AI Overviews grew 34% in six months. Surfer SEO's analysis of
When you aggregate citations across every major AI platform (ChatGPT, Perplexity, Google AI Overviews, Gemini, Copilot, Grok), one domain sits at the top. Not Wikipedia. Not YouTube. Not Forbes. Reddit. Profound analyzed over 4 billion AI citations and found Reddit accounts for 3.11% of total citations across all platforms.
Here's a number that should reshape your editorial calendar as much as it shaped what Radiant Elephant does for our clients: 50% of AI citations come from content less than 13 weeks old. That's not a guideline. That's the effective shelf life for AI
For roughly fifteen years, the SEO industry has treated backlinks as the primary currency of online authority. More links from better sites equal higher rankings. Simple. Measurable. Actionable. Even Radiant Elephant abided by this philosophy. AI search doesn't work that way. Ahrefs studied 75,000 brands in August
Here's a stat that should change how you think about every page on your website, it sure changed the way my SEO agency Radiant Elephant approaches content: 55% of AI Overview citations come from the top 30% of a page's content. If your best data, your
Every SEO blog in 2026 has a GEO guide. Most of them recommend the same generic list: "optimize for AI," "structure your content," "be authoritative." Very few of them cite actual research. And almost none of them tell you which specific tactic produces the
The short answer: you need to be cited as an authoritative source in the training data and real-time retrieval systems ChatGPT uses. That's a layered problem, and most of the "quick fix" advice floating around is garbage. I run an agency, Radiant Elephant, that does
News
This is Rank And Convert, a brand new site by Gabriel that's just getting started. Things will be up and running here shortly, but you can subscribe in the meantime if you'd like to stay up to date and receive emails when new content is published!