Who Gets Cited in AI Search? The Source Concentration Data

Research into AI search citation patterns consistently shows that a small number of domains capture the majority of citations across ChatGPT, Perplexity, and Google AI Overviews. Studies tracking citation distribution in 2025 and early 2026 found that the top 1% of cited domains account for roughly 60% of all citations in many query categories. For most UK businesses, this means getting into the cited set at all is harder than simply having good content.

The core statistic

Analysis published by Profound in Q1 2026, covering over 4 million AI-generated responses across ChatGPT, Perplexity, and Google AI Overviews, found that the top 1% of cited domains received approximately 60% of all source citations. The bottom 50% of cited domains each appeared in fewer than 0.1% of relevant responses.

This is a sharper concentration than traditional organic search, where the top 10 ranking pages share visibility across a much wider range of domains for any given query set.

How the data was collected

Profound's methodology involved querying each AI platform with a standardised set of commercial, informational, and navigational prompts. Responses were parsed for inline citations and footnote references. Domains were then ranked by total citation frequency across all response types.

The dataset covered queries in English, with a weighted mix of B2B and B2C topics. UK-specific queries were a subset of the total. The findings were consistent with earlier work by Semrush and BrightEdge, both of which noted high citation concentration in their own 2025 studies, though each used different query sets and platform mixes.

Concentration by platform

Perplexity shows slightly lower concentration, partly because it surfaces more citations per response and draws on a broader real-time index. ChatGPT and Google AI Overviews are more selective, favouring domains with strong existing authority signals.

What types of sites dominate?

Across all three platforms, the most-cited domains in commercial and informational queries tend to share several characteristics. They are typically large-scale editorial publishers, well-established industry bodies, government or quasi-governmental sources, or sites with very high topical authority in a narrow vertical.

Wikipedia, national news outlets, NHS pages, and major trade associations appear disproportionately often. In B2B software queries, a small cluster of review and comparison platforms, including Gartner, G2, and Capterra, are cited far more frequently than individual vendor sites.

Smaller specialist sites do appear, but usually only when no authoritative large-domain source covers the specific sub-topic. This creates what researchers have called a "citation gap" for niche queries, where AI platforms must reach further down the authority ladder.

### Who Gets Cited in AI Search? The Source Concentration Data: key figures

| Metric | Value | Source |
| --- | --- | --- |
| Share of all AI search citations held by the top 1% of cited domains | ~60% | Profound, Q1 2026 analysis of 4 million AI-generated responses |
| Share of citations held by top 1% of domains on Perplexity | ~55% | Profound, Q1 2026 |
| Share of citations held by top 1% of domains on ChatGPT (browsing) | ~62% | Profound, Q1 2026 |
| Share of citations held by top 1% of domains on Google AI Overviews | ~65% | Profound, Q1 2026 |
| Median citation depth per Perplexity response | 5 sources | Profound, Q1 2026 |
| Median citation depth per ChatGPT browsing response | 3 sources | Profound, Q1 2026 |
| Median citation depth per Google AI Overviews response | 3 sources | Profound, Q1 2026 |
| Share of citations held by bottom 50% of cited domains | fewer than 0.1% each | Profound, Q1 2026 |

Source: Profound, Q1 2026 analysis of 4 million AI-generated responses, Profound, Q1 2026

What this means for UK businesses

The concentration data has direct implications for how businesses should think about AI search investment. Simply publishing accurate content is not sufficient if an AI platform's retrieval layer consistently prioritises a handful of established domains over everyone else.

For businesses in competitive categories, the practical question is not "how do I rank?" but "how do I become a source that authoritative sites cite, quote, or reference?" AI platforms often cite secondary aggregators rather than original sources, which means appearing in Gartner, a trade association report, or a major news outlet can deliver citation value that your own site cannot generate directly.

For businesses in less competitive or highly specialised verticals, the picture is more encouraging. When query volume is low and large publishers have not covered a topic in depth, smaller authoritative sites do break through. This is particularly true in UK-specific regulatory, legal, and sector-specific queries where US-centric sources are less relevant.

The fundamentals of AI search optimisation still apply: structured content, clear entity signals, and technical accessibility all help. But the concentration data suggests businesses should layer a digital PR and third-party citation strategy on top of on-site optimisation rather than treating on-site work as sufficient on its own.

How to act on this data

1. Audit where you are currently cited

Before investing in new content, find out whether your brand appears in AI responses at all, and if so, through which intermediate sources. A free AI visibility audit can establish a baseline quickly. Knowing whether you are cited directly or only via a third party changes the strategy considerably.

2. Prioritise third-party citation signals

If your brand is absent from AI responses, focus first on getting mentioned in sources that AI platforms already trust. Target trade press, industry bodies, and authoritative comparison or review platforms relevant to your sector. A single citation in a Gartner report or a BBC News article may carry more AI search weight than dozens of new pages on your own site.

3. Target the citation gap

Identify sub-topics within your niche where large publishers have thin or outdated coverage. These are the queries where a well-structured, authoritative page on your own domain is most likely to be cited directly. Use AI platforms themselves to test which queries return responses that cite smaller specialist sources rather than major publishers.

4. Review your technical signals

Even if you secure third-party mentions, AI crawlers still need to access and process your site. Review your schema markup and ensure you are not inadvertently blocking the crawlers that matter most. The AI crawler directory lists the key bots and their identifiers.

Comparison: AI citation concentration vs organic search

Organic search has always favoured established domains, but the gap is meaningfully wider in AI-generated responses. This is partly a function of how retrieval-augmented generation works: the model needs high confidence in a source before surfacing it, and established authority is a strong proxy for that confidence.

Limitations of this data

Citation concentration figures vary significantly depending on which query categories are included. A dataset heavy in medical or financial queries will show higher concentration than one focused on niche B2B topics. The Profound dataset is among the larger public studies, but it is still a sample rather than a census of all AI-generated responses.

Platform behaviour also changes as models are updated. Google AI Overviews has adjusted its citation logic several times since launch. Figures that were accurate in Q1 2026 may shift within months.

Frequently asked questions

Does this mean small businesses cannot appear in AI search?

Not at all, but it means on-site content alone is unlikely to be sufficient in competitive categories. Small businesses with genuine topical depth, strong third-party mentions, and good technical signals do appear in AI responses, particularly for niche or regional queries where large publishers have limited coverage.

Are UK-specific queries as concentrated as global ones?

UK-specific queries, particularly those involving UK law, NHS services, or UK-only sectors, tend to show slightly lower concentration because the pool of authoritative UK sources is smaller than the global English-language pool. This makes UK specialist sites more competitive in locally framed queries than they would be in generic international ones.

How quickly can a business improve its citation rate?

Third-party citation campaigns typically take three to six months to show measurable impact on AI search appearances, based on current agency reporting. Direct on-site improvements, such as schema and content structure, can show faster results for niche queries but have limited impact in concentrated categories.

Which AI platform is easiest to break into as a new entrant?

Perplexity currently shows the lowest citation concentration of the major platforms, surfacing more sources per response and refreshing its index in near real time. For businesses starting out, it is often the first platform to test citation visibility improvements against.

Next step

If you want to see where your domain currently stands in AI-generated responses, a free AI visibility audit will show you which platforms are citing you, which queries trigger your appearances, and how your citation rate compares to category benchmarks. You can also review the UK agency comparison if you are looking for specialist support with a third-party citation strategy.