Modern brands know that appearing in AI search is a must.
What they may not understand as well is that AI visibility is more than just being mentioned by name. AI platforms also link to and cite brand content that they've deemed a trustworthy source of information. And sometimes an AI engine's response will include both a brand mention and citation simultaneously.
An analysis of 456 Stacker-distributed stories found that of the three possible types of AI visibility, citations were by far the most common. In fact, AI responses that included a brand citation were more than three times as common as those that mentioned a brand by name. We also found that citation-only responses grew 33% in the four weeks following Stacker distribution.
It's important to understand the differences in how AI visibility manifests as more people move their search activity to these platforms. According to a report from marketing agency Eight Oh Two, a growing number of people — 37% of consumers — start their search for information on AI platforms rather than Google. That same report, which is based on a survey of 500 active AI users, found that 47% of consumers said that LLMs influence their trust of brands.
For brand teams building content with the goal of improving AI visibility, optimizing for citations through high-quality articles (that can then be distributed) is a logical next step. To be cited by AI platforms, content should be original, data-backed, and easy for LLMs to parse.
Methodology
Stacker tracks AI search visibility in partnership with Scrunch, monitoring how six AI platforms (ChatGPT, Claude, Perplexity, Google Gemini, Google AI Overviews, and Google AI Mode) respond to 30 topic-relevant prompts that are unique to each story. The performance of each story is then measured at five different phases: a baseline before distribution, then once each week for four weeks after a story is distributed.
For this particular analysis, we looked at 456 stories that were distributed on the Stacker Newswire between May and July 2026.
There are three ways that a brand could be considered "present" in an AI response:
- Citation: AI links to brand content — either on the brand domain or on a version of the content that was distributed on the Stacker network
- Mention: AI names the brand in its response but doesn't link to it
- Citation + Mention: AI names the brand and links to it
When making our calculations, responses that included both a citation and a mention were added to both buckets.
Citations were more common than mentions
Our analysis found that brand content is being treated as an authoritative source more often than the brand itself is being mentioned by name.
We found that brands were cited (including citation-only responses as well as responses with both a citation and a mention) in 13.6% of responses, compared to the 3.8% of responses that mentioned a brand by name.
We also found that, when looking at all of the story data in aggregate, the number of citation-only responses grew 33% in the four weeks after distribution via the Stacker Newswire, going from 9% of responses to 12% by the fourth week. This indicates that distribution had a measurable effect on citation-based visibility, separate from a brand's existing authority. Total presence grew 25%, going from 12% at the baseline to 15% at the end of the monitoring window.
Mention-only responses (meaning responses that do not include a brand link at all) stayed relatively flat after distribution, indicating that existing brand authority is a larger factor when it comes to earning mentions specifically, compared to citations.
How citations vary by platform
We also found that each of the AI platforms cited varying numbers of sources.
Google AI Mode and Gemini tended to make the most citations, with a median of 12 each. However, Gemini only provided citations in 59% of its responses, and it cited a median of just four unique domains, indicating it tends to rely on the same sources in its responses.
On the other end of the spectrum, ChatGPT provided the fewest number of citations, with just five at the median. However, it included citations in 99% of its responses.
We previously studied how brands’ overall presence varies by platform and found similar results — Google AI Overviews and Google AI Mode were the most friendly to brand content, while ChatGPT was the toughest one to crack.
It’s a reminder that AI visibility does not manifest the same way on every platform, and that each should be judged by its own unique metrics.
What this means for content teams
If brand content being cited is the most common path to AI visibility, it follows that the biggest impact brand teams can have lies in making their content highly citable.
That means creating content that is unique, rich in data, and editorially sound. We’ve found in previous analyses that stories that answer specific questions for consumers and include clear explanations of their methodology tend to perform well from a GEO perspective.
They also tend to earn the trust of high-authority publishers like those in the Stacker network. Recent research has continued to show that earned media — in the form of journalistic articles, academic research, and similar third-party coverage — generates authority signals that are valued by AI. That’s likely to contribute to growing AI visibility.
Madeline Stone is the Content Manager for Editorial & Insights at Stacker. She was previously a longtime business and tech journalist at Business Insider and a content and communications consultant for startups.
