.svg)
Get a Demo
Interested in trying Goodie? fill out this form and we'll be in touch with you.
Oops! Something went wrong while submitting the form.
.svg)
.svg)
A few websites, most notably Wikipedia, YouTube, Reddit, Quora, and Reuters, account for a notable percentage of the citations that LLMs attach to their answers. Since these citations shape what people read when they consult AI tools, domains that dominate “citation share” reap disproportionate visibility, authority, and downstream traffic.
Borrowing a term from SEO, “citation share” is the new PageRank.
The domains most frequently cited by LLMs are reshaping the flow of online authority. Based on an analysis of 5.7 million citations collected between February and June 2025, here are the 20 most cited domains in AI answers across Gemini, ChatGPT, Claude, and Perplexity:
These are the top 20 domains (we could only fit so many on the page), but you can download the full list of 50 here. These domains that are most frequently cited in LLMs can then be categorized by how susceptible they are to marketing influence:
The highly influential brands are social media platforms or sites made up of user-generated content, business and commerce sites, corporate sites, and reference pages.
Analyzing the first category reveals that social is the dominant lever in shaping LLM citations and marketing influence by extension. For CMOs, this reinforces that social platforms set the baseline for how AI systems perceive brand authority. This means that investments in social listening, influencer partnerships, and community engagement can shape AI visibility.
Moderately influencable websites include news outlets, the content of which can be directly influenced through PR and press releases such as reuters.com, apnews.com, cnn.com, nytimes.com, yahoo.com, axios.com, washingtonpost.com, wsj.com, foxnews.com, theguardian.com, bbc.com, latimes.com, theatlantic.com, and aljazeera.com.
Low and no marketing influence describes sources that are academic or scientific, technical, or medical.
Combining the data for sites that are considered to be high and moderate influence sources reveals that 74% of the most-cited domains in LLMs are susceptible to marketing influence. For brands, this uncovers a significant opportunity to shape how AI systems learn about and represent their companies.
Domains that are the “most cited” appear most frequently in footnotes, carousel cards, sidebars, or inline links that answer engines attach to their summaries. Citation share is more granular than traffic or backlinks (as was the case with traditional SEO), and reveals which brands are trusted most consistently by AI systems as sources of truth. It is essential to position your brand as a source of truth in order to improve the frequency and sentiment with which LLMs talk about your brand.
Recent analyses also reveal that AI answer engines concentrate citations on a few brands. For example, Google’s AI Overviews cite news sources in about 20% of responses, and within those, the top 10 news outlets capture nearly 80% of all mentions, with outlets like BBC, The New York Times, and CNN dominating.
This highlights a distinction between mention share (how often a domain is cited) and impression share (citation frequency adjusted by visibility context). A source might have few citations but still have a high impression share if it appears in places that garner attention in AI summaries.
For example, a brand can have thousands of LLM citations, giving it a high citation share, with only a select few being for key phrases or queries that have high visibility within the topic. This concept and contextualization make both mention share and impression share valuable AI visibility metrics to track.
Brands in the top quartile for web mentions receive over 10x more citations in AI Overviews than those in the subsequent quartile; and with more citations comes more highly qualified and converting traffic. To capture this high-converting traffic, brands should improve their presence on LLMs. One way to do this is by leveraging answer engine optimization platforms like Goodie AI, which can help you own your narrative on AI platforms.
Showing up in AI summaries is important because user adoption of these summaries is rapidly expanding. The majority (58%) of respondents in a March 2025 study visited a search result page with a summary generated by AI.
With increased visibility in AI, brands become more authoritative. More authoritative brands are cited more frequently, and more citations compound a domain’s authority and visibility. This creates a virtuous cycle for the dominant few, while the others are left behind. From a CMO’s perspective, the takeaway is that citation concentration signals where to direct budget (toward PR, data partnerships, and earned authority plays, since these create compounding effects in AI funnels.
Becoming the top search result is essential for your brand to stand out against the competitors as a result of Google’s “winner takes all” attitude with AI Overview. This is similar to the importance of being the top result in SEO. LLM citations carry similar authority to an AI Overview, but Google’s placement of it at the top of search results makes it far more visible.
Research and analytics vendors track LLMs and their citation patterns. There are three primary categories of these sources: direct scraping, crawling at scale, and benchmarking studies.
Google and OpenAI display citations in their UI for transparency. Researchers can capture them by scraping or parsing the UI to analyze the AI citation behavior. By tracking which domains are cited most frequently on a large scale, this approach enables the automated parsing of the response UI by query.
Specialized tools and vendor panels scan AI responses for outbound links and domains. Dedicated platforms such as Goodie continuously monitor AI answer engines by logging outbound links and domain mentions over time. These panels provide structured datasets across multiple AI models, allowing for comparisons across answer engines of citation share and trust patterns.
Industry research compares citation sets from AI models with traditional SEO APIs to analyze overlap and identify unique biases present in AI selection criteria.
Citation share, like any metric, can fluctuate over time and across different AI tools. To verify accuracy, researchers can sample queries at various times of day to account for fluctuations, as citations may vary by hour. Cross-referencing multiple tracking vendors can smooth out sampling noise. Manually inspect queries that matter the most, since automated scraping can have an error margin.
To find the most cited domains in AI search, consult comprehensive analyses created from AI model outputs and public citation domains. If this is something you’re looking for, you’re in the right place; this report is one of them!
Our analysis of 5.7 million citations reveals that Wikipedia, Reddit, and Reuters top the list of the 50 most influential domains in responses generated by leading AI models. Wikipedia leads by volume and influence, which reflects its ubiquitous utility for facts, context, and foundational knowledge. Reddit follows and represents the importance of discussion and user-generated insights, but commercial intent for conversions tends to rely on more niche review sites, forums, and affiliates.
News publishers like Reuters, Forbes, AP News, and The New York Times dominate due to their global reach and commitment to updating journalism. Social and UGC domains such as LinkedIn, YouTube, Quora, X, and GitHub form the personalized information component often written by peers. These domains demonstrate that AI models value the curated authority of publishers and the authenticity of collective knowledge platforms.
Commercial domains maintain their influence, with Amazon and Apple often cited in commerce and product questions. Academia and science are well represented by Nature, Science.org, and major university outlets, which indicates that research sources are valued in technical or complex queries. Specialized reference sources like Healthline and Investopedia are also prominent within their niches.
Across the 50 domains, nearly 3 million citations are concentrated, while the remaining millions are distributed among the 40,000+ less frequently referenced sites. This underscores the sharp drop-off in domain authority outside of the top tier.
Yes! There are multiple platforms that help you improve your AI visibility. This practice is often called Answer Engine Optimization (AEO), Generative Engine Optimization (GEO), or Large Language Model Optimization (LLMO).
To find the tool that suits your AEO needs, check out our recent blog post evaluating the best AEO software tools.
Before you settle on a tool to use, it’s important to understand the tactics that are used to improve citation share. This requires a combination of strategic PR, technical optimization, and content reorientation.
Platforms like Goodie serve as end-to-end AEO hubs with strategies for providing insights and recommendations, and implementing improvements, consolidated in one location.
Citation preferences differ between LLMs. Underlying training data, retrieval mechanisms, and alignment with user intent all shape these preferences. Let’s take a look at three industries where domain influence varies depending on these factors:
Wikipedia is ubiquitous across all industries; Fintech included. Reddit and Investopedia also increasingly surface for financial queries, particularly during periods of financial volatility as users seek live insights verified by other users. Further, Reddit is important for awareness, research, and early-stage validation.
For more information, we recently did a finance-specific breakdown of the most cited domains across ChatGPT, Perplexity, Claude, and Gemini.
In this industry, information-seeking tends to be incredibly niche and driven by complex technical queries. As a result, data-driven guides that also focus on integrations dominate citations, especially if they live on educational blog platforms or technical forums.
For more information, we recently did a similar breakdown for the top cited domains in B2B SaaS across ChatGPT, Perplexity, Claude, and Gemini.
Domains that are dominated by influencers and social networks are the most influential in beauty and personal care. Consumer review sites, for example, hold outsized influence as users gravitate towards authentic content with peer testimonials and product discovery with social trends.
Across the four major AI models, NerdWallet emerges as the singular leader across models and appears for top affiliates and listicles. This signals that both AI models and users recognize the site’s authority in comparative guidance for banking, payments, and personal finance. News outlets such as CNBC, Forbes, Bloomberg, and Business Insider hold strong as references, especially for breaking news, market trends, and analysis of financial products. Reference services like Investopedia and Bankrate surface when users seek education, definitions, or methodological clarity.
Each AI platform has unique citation characteristics:
Across all platforms, the influence scores smooth out gaps from raw citation numbers and present a holistic picture of digital authority in real buying scenarios. Financial decisions are therefore most likely to be shaped by a mixture of practical reviews, authoritative journalism, and reference points that reinforce which digital brands companies should prioritize for partnership, content placement, and reputation management.
Affiliate listicles, user-generated content, and authoritative industry players drive the B2B SaaS citation landscape. Curated software rankings on places like G2, Capterra, and PCMag have significant influence in the vendor discovery phase.
Peer validation on places like Reddit, Wikipedia, and TrustRadius offers feedback from the real world and experimental insights that can push users toward action. These conversations contribute to early-stage awareness and credibility building. Wikipedia and TrustRadius are more influential for commercial intent, whereas Reddit is more important for early-stage crowdsourcing and validation.
Reddit still plays a role in surfacing authentic consumer conversations, but when it comes to purchase queries, LLMs lean on retailers and affiliate review hubs where structured content maps directly to buying intent. The beauty and personal care landscape is dominated by social platforms, editorial authority, and major retail marketplaces. Reddit, Sephora, Allure, Byrdie and similar domains combine user discussion, editorial storytelling and direct commerce to drive discovery and conversion.
Retailers like Sephora, Amazon, and Ulta surface as sales channels and influential content hubs where product collections and reviews shape purchase intent. Visibility across these types is table stakes for influencing the customer journey.
For brands looking to maximize their impact during the AI search revolution, improving domain influence is priority number one. Here’s how to do it:
Once the above tactics are applied, measuring success and course-adjusting becomes critical. And when it comes to the right tool for the job, Goodie’s got your back. Get in touch with us today to get started.
.svg)
.svg)
.svg)
.svg)
