Who Gets Cited in AI Search? The Top Source Statistic

Research published in 2025 by Seer Interactive found that roughly 84% of ChatGPT citations link to pages that also rank in Google's top 10 for related queries. That overlap is the single most important data point for any UK business planning an AI search strategy: existing search authority still heavily predicts AI citation likelihood, but it is not the whole picture.

The headline figure

Seer Interactive analysed thousands of ChatGPT responses across a range of informational queries and matched the cited URLs against Google Search rankings. Approximately 84% of cited pages held a position in Google's top 10 for a closely related keyword at the time of analysis.

That single figure has significant implications. It suggests that AI platforms, at least for ChatGPT, are still leaning heavily on the same corpus of content that traditional search engines surface. If your pages are not ranking well in Google, your chances of appearing in a ChatGPT citation are materially lower.

How the data was collected

Seer Interactive ran a structured sample of informational queries through ChatGPT and recorded every URL cited in the model's responses. Each URL was then checked against Google Search results using keyword mapping to find the closest matching query. A page was classed as a "top 10 result" if it appeared on the first page of Google for at least one closely related term.

The study focused on ChatGPT with web browsing enabled. It does not directly cover Perplexity, Gemini, or Google AI Overviews, which each have different retrieval mechanisms and index freshness characteristics. The findings should therefore be treated as directional rather than universal.

Comparison across platforms

The absence of comparable published figures for Perplexity and Claude means direct cross-platform comparison is not yet possible at the same confidence level. If you are building a multi-platform strategy, it is worth reviewing the AI search statistics hub for the latest data as it emerges.

What types of content are cited most often

Beyond the Google-ranking correlation, the Seer Interactive data and supporting research from SparkToro and others point to consistent patterns in the types of pages that attract citations:

• Definitional and explainer content ("what is X", "how does X work")
• Structured comparison pages with clear headings and data tables
• Pages with clear authorship signals and publication dates
• Content hosted on domains with established topical authority
• Pages that are crawlable and not restricted by robots.txt directives

The crawlability point is worth emphasising. If your robots.txt file is blocking AI crawlers, even high-quality content will not be indexed. The robots.txt guide on SEOCompare covers exactly which directives affect different AI platforms.

### Who Gets Cited in AI Search? The Top Source Statistic: key figures

| Metric | Value | Source |
| --- | --- | --- |
| Percentage of ChatGPT citations that also rank in Google's top 10 for related queries | ~84% | Seer Interactive, 2025 |

Source: Seer Interactive, 2025

What this means for UK businesses

The 84% overlap figure cuts both ways. On one hand, it is reassuring: if you have already invested in traditional SEO and you are ranking well in Google, you have a strong foundation for AI citation visibility. You are not starting from scratch.

On the other hand, 16% of cited pages do not appear in Google's top 10. That gap represents content that AI platforms are pulling from elsewhere, whether that is strong off-site mentions, high-authority domains that rank lower for specific terms, or content that is simply well-structured for machine parsing even if it lacks traditional ranking signals.

For businesses in competitive sectors, that 16% is where differentiated AI search optimisation work happens. Structured data, schema markup, clear entity associations, and strong third-party mentions all contribute to citations that go beyond pure Google ranking correlation. The schema and AI visibility guide explains which schema types have the most measurable impact.

Practical implications by business type

How to act on this data

The most direct action is to identify the pages on your site that are closest to ranking in Google's top 10 for your target queries. These are your best candidates for AI citation with incremental work.

For each candidate page, check three things. First, confirm that AI crawlers can access the page. Second, verify that the page has appropriate structured data. Third, look at whether the page is cited in external sources such as industry directories, press coverage, or authoritative third-party sites.

If you are unsure where to start, the free AI visibility audit can identify which of your pages are currently being picked up by AI platforms and where the gaps are.

For businesses weighing whether to handle this in-house or work with a specialist, the agency comparison on SEOCompare lists UK providers with verified AI search track records.

Frequently asked questions

Does this mean traditional SEO and AI search optimisation are the same thing?

Not entirely. The 84% overlap shows that Google ranking is a strong predictor of AI citation, but the two disciplines diverge on structured data, entity optimisation, and off-site citation signals. Treating them as identical would mean missing the 16% of citations that come from outside the top 10.

Is the 84% figure applicable to Perplexity and Gemini?

The Seer Interactive study focused on ChatGPT with web browsing enabled. Perplexity uses its own crawler alongside Bing, and Gemini draws from the Google index, so similar correlations are plausible but have not been confirmed at the same scale. Treat the figure as directional for other platforms.

If I rank in Google's top 10, will I automatically appear in AI citations?

No. Ranking in the top 10 improves your probability significantly, but AI platforms also filter based on content relevance, crawlability, page structure, and query intent matching. Technical barriers such as blocked crawlers or missing schema can prevent citation even for well-ranked pages.

How often does this data change?

AI platform retrieval behaviour evolves with model updates and index changes. The 84% figure is from a 2025 study and the precise percentage may shift as platforms like ChatGPT and Perplexity update their retrieval stacks. Monitoring your own citation data regularly is more reliable than relying on any single study over time.