Neil Patel’s Research Reveals Shift in AI Citation Sources

In a recent LinkedIn post, Neil Patel discusses significant shifts he’s observed in how AI models, particularly ChatGPT, cite their sources, suggesting a potential re-evaluation of traditional SEO strategies.

Patel’s research, which analyzed 4,308 prompts and 1,161 citations, highlights a dramatic decrease in AI’s reliance on top-ranking Google results. He notes that the percentage of ChatGPT citations drawn from Google’s top ten results has fallen from 76% to 38%.

Google’s top ten results used to account for 76% of ChatGPT citations. That number has dropped to 38%, and 90% of the pages AI cites now rank 21 or lower on Google.

This finding challenges the long-held assumption that strong Google rankings automatically translate into visibility within AI-generated answers. According to Patel, this indicates a diverging landscape where traditional SEO may not be as effective for AI search.

The Declining Influence of Top-Ranked Pages

Patel’s analysis points to a growing trend where AI models are increasingly referencing content from much lower on the Google search engine results pages (SERPs). He states that a substantial 90% of the pages cited by AI in his study now rank outside the top 20 on Google.

This suggests that AI’s underlying algorithms for source selection might be prioritizing different factors than those that drive traditional search rankings. As Patel puts it:

If you’re investing in SEO and assuming it carries over into AI search automatically, it does not.

Key Patterns in AI Citations

The research identified several patterns influencing which content gets cited by AI. While not detailing all five patterns from his accompanying video, Patel emphasizes that factors beyond mere keyword optimization are at play.

Factors Beyond Traditional SEO

Patel’s work implies that site structure, content freshness, and entity association are becoming increasingly important for AI discoverability. Furthermore, he points out that platform differences exist, with AI models like Gemini, Claude, and Meta AI potentially exhibiting distinct citation behaviors compared to ChatGPT.

In Neil Patel’s view, this necessitates a strategic shift for businesses and content creators. The focus may need to broaden from solely optimizing for Google’s ranking factors to understanding how AI models process and prioritize information.

In this video, I break down the 5 patterns we found across 4,308 prompts and 1,161 citations, covering everything from site structure and content freshness to entity association and platform differences across ChatGPT, Gemini, Claude, and Meta AI.

Ultimately, Patel’s insights suggest that the future of online visibility may require a more nuanced approach, one that accounts for the evolving ways AI consumes and presents information, potentially decoupling it from traditional search engine dominance.