In today’s evolving AI landscape, tools like ChatGPT, Gemini, and Claude are key sources of information. But what do these chatbots actually “read” and cite? For PR professionals, this insight is crucial for shaping narratives and visibility.
A new MuckRack report analyzed over a million AI responses from July 2025, revealing how generative AI citations differ from traditional SEO and why PR professionals need to rethink their digital strategies with Generative Engine Optimization (GEO) in mind.
The Role of Citations in AI-Driven Insights
A key finding from Muck Rack’s report is that AI models like ChatGPT provide more accurate, current answers when they access real-time citations. Instead of outdated info, they draw from trusted, up-to-date sources like news articles and blogs.
For example, asked “Who’s the worst MLB team?”, ChatGPT first named the 1962 Mets but updated to the 2024 White Sox with citations enabled, highlighting how AI relies on current published content and why earned media is increasingly vital.
What Are AI Models Citing?
Muck Rack found that AI strongly favors earned media – credible, third-party content – over paid or promotional materials. The report found that:
- 95% of AI-cited links are non-paid media.
- 27% come from journalistic outlets.
- 37% come from third-party corporate blogs and content (not owned by the company targeted in the inquiry).
- Only 9% comes from owned corporate accounts.
- Just 1% comes from press releases.
This data shows that AI models prioritize trustworthy journalism, expert blogs, and research over marketing-driven sources.
Question Type Matters
AI models cite different sources based on how the question is framed. For example:
- Advice and opinion seeking → corporate blogs and thought leadership.
- Current events → journalism (about 50% of the time).
- Fact-checking → Wikipedia, academic, and government sources.
Interestingly, different AI models show different behaviors. Claude tends to cite academic and government sites more and journalism less, while ChatGPT favors mainstream news.
Citation Patterns by Industry
AI citations also vary by industry. For instance:
- Healthcare responses cite government and NGO content more than average.
- Finance and media cite journalism most heavily.
- Energy and hospitality industries cite more owned corporate content.
This means PR teams need to understand the citation “ecosystem” within their industry to optimize for AI visibility.
What Is Generative Engine Optimization (GEO)?
The rise of AI-powered search is making way for a new practice called Generative Engine Optimization; an evolution of SEO tailored for AI-native platforms.
Unlike traditional SEO, which relies on keywords and backlinks, GEO focuses on creating credible, timely content AI trusts, securing earned media coverage, producing clear thought leadership, and using multimedia with transcripts to boost visibility.
What’s Next for Businesses?
To stay visible in AI-driven search results, businesses should prioritize earned media, working with a PR team to pitch credible stories and expert insights to trusted outlets. Creating helpful, non-promotional owned content (like blogs and FAQs) will also increase AI citation potential.
Understanding which sources AI favors in your industry and using GEO tools to monitor and refine your brand’s AI presence are key to staying ahead.
Why PR Must Evolve with AI
As AI chatbots become a more important gateway for information, PR must evolve to stay relevant. Embracing Generative Engine Optimization alongside SEO ensures messages reach audiences, build trust, and keep organizations visible in an AI-driven world.