Close Ranks: Study Shows Top-Ranked SEO Results More Likely To Appear As AI Sources

SEO has been in the crosshairs and catching strays from the rapid rise of AI quite a bit recently, and while generative AI will no doubt change Search and content optimization as we currently know it, SEO itself is far from dead.

In fact, a recent ZipTie analysis of 25,000 queries has “discovered something crucial that contradicts the popular narrative,” according to co-founder Tomasz Rudzki: that websites that rank #1 in Google’s traditional search results have a 25% chance of appearing as the source in AI Overviews, and likely other AI platforms as well.

“The higher you rank in Google’s top 10, the more likely you are to appear in AI search results across platforms,” writes Rudzki. “This isn’t speculation — it’s based on real queries from real users across ChatGPT, Perplexity, and Google AI Overviews.”

(Source: ZipTie)

Through his analysis and internal Google documents (some of which came to light only from Google’s antitrust trial), Rudzki theorizes that AI Overviews identifies the top query-matching documents, pulling the relevant information, then synthesizes a coherent answer through Gemini. 

“This explains exactly why traditional Google rankings remain so crucial for AI search inclusion,” he writes, noting that factors such as user-personalization, caching, and Google’s “query fan-out technique” of developing a response might be the reason some sources outside the top ranks might still appear in AI’s results.

As Search Engine Journal’s Matt G. Southern points out, that 25% hit rate by top-ranked content means 75% are missing out, creating “an opportunity” for those who adapt.

“In my experience, to be successful in AI Search, you should optimize for traditional search results [and] make sure your content is relevant,” writes Rudzki. “SEO isn’t dead — it’s evolving rapidly. The fundamentals still matter, but execution must shift from keyword optimization to user-intent optimization. Instead of asking ‘How do I rank higher?’ start asking ‘How do I better serve users who have specific questions?’”