We Tracked AI Referral Traffic for a Year. Here's What the Data Actually Tells Us About SEO and AEO.

In late 2024, a routine analytics review flagged something unusual. Referral traffic from ChatGPT. Then Perplexity. Then Gemini. Not one-off visits. A consistent, growing pattern across two climate and energy media publications we manage, operating across 13 Asian markets.

By Q4 2024, AI referral traffic had almost tripled compared to the previous quarter across both sites. We had built custom GA4 channel groupings specifically to track AI sources, and a dedicated dashboard to monitor them. The data was unambiguous: something was shifting.

That triggered a real strategic question. Should we invest more in AEO-optimized content, the kind structured specifically to be cited by AI platforms? Or should we double down on the SEO work we were already doing? We decided to find out.

The Global Picture

What the Industry Data Says About AI Referral Traffic

Our internal observation wasn't an anomaly. The broader data confirms that AI referral traffic is growing fast globally, even if it's still small in absolute terms.

AI referral traffic grew from just 0.02% of total web traffic in 2024 to approximately 1% by 2025, a sevenfold increase in a single year.^[1] Adobe Analytics reported that AI platforms generated 1.13 billion referral visits globally in June 2025 alone, a 357% increase from June 2024.^[2] An analysis of 2.3 billion sessions found AI traffic grew 796% from 2024 to 2025.^[3]

ChatGPT dominates the AI referral landscape with approximately 77–87% of all AI-driven clicks depending on the study.^[1][4] Perplexity holds around 15% globally. Gemini trails at 6.4%, though it's growing the fastest in percentage terms. Worth noting: by March 2026, ChatGPT's share had already fallen to 56.72% as Gemini surged to 25.46%.^[5] The platform distribution is shifting faster than most marketers realize.

The quality signal matters as much as the volume. AI-referred visitors convert approximately 1.2x higher than organic search visitors across all channels.^[3] One B2B study found ChatGPT converting at 15.9% versus Google Organic at 1.76%, a nearly tenfold difference.^[6] ChatGPT visitors also view an average of 2.3 pages per session versus 1.2 for organic search.^[6] This is not a novelty channel. It is a high-intent one.

The Analysis

Our Hypothesis and How We Tested It

The question we set out to answer was specific: does a lower Average Position in Google Search (i.e., better SEO ranking) lead to more views from AI referral traffic?

To test it, we and our data team analyzed website analytics and Google Search Console data across both publications. We correlated Average Position against Views from AI referral traffic at the page level. We ran Pearson correlation analysis across the full dataset, then segmented by publication and by geography to look for patterns within the data.

A note on limitations. At the time of this analysis, AI referral traffic accounted for less than 1% of total organic traffic on both sites. That means the dataset of pages with meaningful AI referral visits is small, which limits the statistical power of any correlation analysis. These findings should be treated as early directional signals, not definitive conclusions. We intend to revisit this analysis as the dataset grows.

The Findings

What We Found

The short answer: the hypothesis is partially true, but the effect is not strong across the full dataset.

The Pearson correlation between Average Position and AI referral views across the combined dataset was -0.2199. Negative because a lower position number means a better ranking. The direction is correct: better SEO ranking does associate with more AI referral traffic. But the effect is weak, and the data is noisy.

The more interesting finding is what happens when you split the data by publication.

Publication A, which has a domain authority of 55 on Ahrefs, shows a meaningfully stronger correlation between SEO ranking and AI referral views than Publication B, which has a domain authority of approximately 20. The higher DA site shows better ranking more reliably translating into higher AI traffic.

This finding points to something the broader research also supports: domain authority appears to amplify the relationship between SEO ranking and AI traffic. It's not just about where a specific page ranks. It's about the overall authority of the domain it sits on. A well-ranked page on a high-authority domain is more likely to earn AI citations than the same page on a lower-authority domain.

External research supports this direction. Ahrefs found that sites with over 32,000 referring domains are 3.5x more likely to be cited by ChatGPT.^[8] Branded web mentions show the strongest correlation with AI Overview appearances at 0.664, significantly higher than backlinks at 0.218.^[9]

By geography, Philippines and Japan content showed the strongest correlation between ranking and AI referral views. Philippines and Malaysia-focused content dominated the top 10 AI traffic-generating pages on Publication A. Country-specific, data-backed content covering these markets consistently outperformed generic global overviews, a pattern that held across both AI platforms and organic search.

Common features of the highest-performing pages for AI referral traffic: a strong 1–2 sentence introductory summary that clearly states the topic or problem; clear subheadings making content easy to parse; a problem-solution content structure with the problem identified early and solutions in separate sections; and fact-based writing anchored to authoritative sources with citations. These structural characteristics appeared consistently across the pages generating the most AI referral views.

Google's Official Position: And Its Limits

What Google's Guide Says, and Where It Falls Short

In May 2026, Google published its official guide to optimizing for generative AI features on Google Search. It's worth reading carefully, both for what it confirms and for what it doesn't cover.

Google's core position: AEO and GEO are still SEO. Their AI features, including AI Overviews and AI Mode, are built on top of their core Search ranking systems using Retrieval-Augmented Generation (RAG). Content is retrieved from their Search index, which means if you rank well in Google Search, you're already better positioned to appear in Google's AI features.^[10]

Google explicitly tells publishers to ignore several tactics circulating in the AEO/GEO space: llms.txt files, "chunking" content for AI, rewriting content specifically for AI systems, and overfocusing on structured data. Their recommendation is to focus on unique non-commodity content, clear structure, good technical SEO fundamentals, and helpful people-first writing.^[10]

This validates our internal finding from Google's own side. Their AI features pull from the same index that SEO ranking determines. Of course better-ranked content gets cited more by Google's AI. That's by design.

The external research bears this out. ChatGPT cites pages from position 21 and below approximately 90% of the time, which is the opposite of what Google's RAG-based system would suggest.^[11] Google AI Overview citations pulled from top-10 organic results dropped from 76% in mid-2025 to 38% by early 2026, suggesting even Google's own AI features are moving away from pure SEO ranking logic as they evolve.^[12]

The honest conclusion is that the relationship between SEO and AI citation is real but platform-specific, evolving fast, and more complex than any single guide can capture. Google's guide is authoritative for Google. It is not authoritative for the broader AI referral landscape.

The Strategic Answer

So, SEO or AEO? Our Answer.

The framing of SEO versus AEO is a false choice, at least for now. The evidence from our analysis, from Google's own documentation, and from the broader research all points in the same direction: SEO is the foundation, with AEO-friendly content structure built on top.

Based on our analysis, here are the practical recommendations:

1. Double down on SEO fundamentals. Ranking well remains the most reliable path to AI visibility, particularly on Google's AI features. Technical SEO, content quality, and crawlability are not optional foundations; they're directly tied to AI discoverability.

2. Build domain authority through high-quality backlinks. Our data showed domain authority amplifying the SEO-to-AI-traffic effect. External research confirms sites with high referring domain counts are significantly more likely to be cited by ChatGPT. Authority building is not separate from AEO strategy; it is AEO strategy.

3. Adopt a localized content strategy for high-performing geographic markets. Country-specific, specific-topic content consistently outperformed broad generic coverage in AI referral performance. The Philippines and Malaysia examples are not anomalies; they reflect a broader pattern where specificity and localization earn more AI citations.

4. Develop AI-friendly content guidelines for your writers. Strong introductory summary, clear subheadings, problem-solution structure, and authoritative citations are structural choices that cost nothing to implement and consistently appeared in the highest-performing pages.

5. Start tracking AI referral traffic in GA4 now if you aren't already. Create a custom channel grouping for AI sources. Build a dedicated tracking view. The organizations that have 12 months of clean data will have a significant analytical advantage over those starting from scratch when this channel becomes more material.

Conclusion

This Is an Early Signal, Not a Final Answer

AI traffic is still under 1% of organic traffic on both publications we studied. The dataset is limited. The platform landscape is shifting faster than any single study can capture. The correlation we found is weak. We will revisit this.

But the direction is consistent. Better SEO ranking correlates with more AI referral traffic. Higher domain authority amplifies that effect. Specific, localized, well-structured, authoritative content earns citations across multiple AI platforms. And the organizations building these foundations now are accumulating an early advantage in a channel that's growing 165x faster than organic search.