Ahrefs’ latest research has reshaped what marketers thought they knew about off-page SEO. New data from the company’s Brand Radar tool shows that brand mentions across the web are now the strongest indicator of visibility in AI search results, outperforming traditional signals such as backlinks and domain rating.
In its analysis of thousands of domains, Ahrefs found that the correlation between branded web mentions and AI search visibility reached 0.664, which is significantly higher than the 0.527 correlation for branded anchors or 0.326 for domain rating. Backlinks (a long-standing SEO benchmark) showed only a 0.218 correlation, suggesting that AI systems weigh mentions more heavily than link structures when determining which sources to cite.
That finding underscores a structural change in how information moves through search. Instead of rewarding link-based authority, AI-driven search systems are recognizing how widely and where a brand is discussed online. Mentions across review sites, discussion boards, and media coverage are becoming the new visibility signals.
Ahrefs’ podcast team, led by CMO Tim Soulo and Director of Content Ryan Law, described this shift as the “era of off-page SEO.” They explained that AI models draw knowledge from two overlapping sources. One is their training data, the massive web snapshots used to build model knowledge, and the other is retrieval data, often called RAG, which supplies live information when a chatbot looks for recent content. A brand mentioned frequently in both layers stands a far better chance of being surfaced during an AI search.
The research also revealed how timing and freshness of content affect citations. When Ahrefs compared millions of URLs referenced by AI assistants such as ChatGPT, Copilot, and Gemini, it found these systems favored newer content far more than Google’s organic results did. Pages updated recently were several times more likely to appear in AI-generated responses, showing that static archives no longer hold the same weight.
Equally important is context. Mentions connected to relevant topics or industry categories strengthen how AI interprets brand identity. A company that appears in conversations about related tools, competitors, or research gains a network of associative signals that shape how models describe it. Mentions in places like Reddit discussions, Quora threads, G2 reviews, or YouTube transcripts are proving especially influential. These platforms supply human language patterns that large models draw from when forming answers.
The study’s insight goes beyond marketing tactics. It hints at how AI itself understands reputation. When a name shows up repeatedly on authoritative domains, in guides, or in credible reviews, the system starts to recognize it as a reliable entity. Visibility then compounds across AI overviews and answer engines.
Ahrefs also found unusual behavior in the data it collected from live model interactions. Around 4 percent of visits from AI sources landed on URLs that didn’t exist, a phenomenon known as hallucinated pages. These phantom addresses (created when models invent links based on assumed site structures) highlight how loosely current AI crawlers map the web. Ahrefs’ analytics now flag such traffic to help site owners correct or monitor how AI perceives their structure.
The same dataset revealed that AI assistants prefer certain types of content. They tend to direct users toward blogs, comparison articles, help pages, research papers, and even indexed PDFs, formats that present clear, factual text. Interactive tools and e-commerce listings, by contrast, rarely appear in AI search traffic because chat interfaces can’t process or display them effectively.
Technical elements also play a part. Pages that rely heavily on JavaScript rendering often go unseen by AI crawlers, which still lag behind Google’s web crawlers in interpreting dynamic content. Keeping critical pages lightweight, well-structured, and readable improves the chance of being captured by AI datasets.
The podcast touched on another developing problem: content spam designed to exploit AI ingestion. Some publishers have begun flooding the web with long, machine-written pages packed with markdown text and no human value. While such material sometimes gets picked up by AI tools, it risks alienating readers and harming brand credibility once clicked. Ahrefs advised brands to focus instead on readable, data-backed writing that benefits both humans and machines.
A more constructive strategy involves tracking where AI assistants cite competitors but ignore your brand. Ahrefs’ Brand Radar tool allows comparison between brands to reveal “mention gaps” — topics or entities where one company dominates the conversation. Filling those gaps with relevant off-site content can realign how AI models represent a brand.
Even as AI search grows, Ahrefs cautioned that Google remains the dominant traffic source. Users may turn to AI assistants for summaries or quick comparisons but still rely on Google for deeper research. Declining traffic to some sites is often caused not by user migration to chatbots but by Google’s own AI overviews, which now occupy key space in results.
The broader conclusion is clear. Off-page SEO has expanded beyond link-building into a wider system of brand presence. Being cited, discussed, and associated with credible sources is now the foundation of visibility in both AI and traditional search. The conversation about your brand (spread across forums, reviews, and media) is quickly becoming the new algorithm.
Notes: This post was edited/created using GenAI tools.
Read next: ChatGPT, Gemini, and DeepSeek Still Confuse Belief with Fact, Study Warns
