Ahrefs Study Debunks Nine AI Search Myths with 15 Million Data Points

Image credit: Search Engine Journal

Ahrefs, a global search engine optimization firm, reported findings from a study of 15 million data points Tuesday, debunking nine prevalent myths surrounding AI search visibility and its impact on SEO strategies.

The research challenges common assumptions about how content achieves visibility within AI models like ChatGPT, Gemini, Perplexity, and Copilot, providing insights for digital marketers.

One key finding indicated that “best-of” listicles often boost the AI visibility of competing brands more than the brand publishing the article itself. AI models frequently use such articles as sources but rarely mention the self-promoting brand, according to Ahrefs.

The study also revealed that 97 percent of llms.txt files receive no fetches from bots or humans, rendering them largely ineffective for influencing AI visibility, Ahrefs reported.

Glen Allsopp, a researcher involved in the study, noted that adding schema markup does not provide an immediate shortcut to AI citations. He explained its impact on AI understanding is slow and indirect, likely operating through knowledge graphs such as Wikidata and Crunchbase.

Mateusz Makosiewicz, another Ahrefs researcher, stated that ranking in classic search remains the primary method for content to be cited by ChatGPT. He added that 88.46 percent of citations originated from the general search index.

However, the research also challenged the assumption that a page one ranking in classic search guarantees AI visibility. Ahrefs found that top search rankings do not directly translate into AI citations.

Lily Ray, a prominent SEO expert, commented on the findings, emphasizing the evolving relationship between traditional search rankings and AI’s content consumption.

The report underscores the necessity for SEO professionals to adapt their strategies based on data-driven insights rather than relying on unproven theories regarding AI search.

Source: Search Engine Journal