Introduction

The media landscape is constantly changing – that’s no surprise. After all, the media landscape has been in flux since the first newspaper was published by Johann Carolus in the 17th century. 

What is unique in this moment, however, is the velocity of change:

• Around 70% of people now get their news from social media, and news organizations are now dealing with “zero-click searches” – when users get all the information they need from AI Overviews without visiting original sources or links.
• AI Overviews take up almost half of a Google search page, and 69% of searches are now zero-click.

Given these significant headwinds, you might be wondering: Does news still matter? 

The answer is a resounding yes. I’ll explain why – along with what PR practitioners can do to take advantage of this environment – in this blog post outlining a recent Fullintel-UConn research report on earned media in the age of AI. 

Why News Matters More Than Ever in the Age of AI

News doesn’t just still matter, according to the research – it may actually matter more than ever. That’s because of its heavily weighted influence in AI search citations and the ability of AI to amplify content appearing on news websites in ways that just weren’t possible before. 

Heavily weighted influence: AI systems treat journalism from news outlets with established, strong reputations as a proxy for credibility and recency. That means news articles are heavily weighted in AI search.

Amplification of quality coverage: Quality news coverage has found an unexpected amplifier in AI search systems. Rather than competing for a single click on a search results page, a well-reported story from a trusted outlet can now be cited across countless AI-generated responses.

The challenge for PR professionals is how to navigate this new, constantly-changing environment to maximize the outcomes of earned media.

Which Content Characteristics Drive AI Engine Outputs?

Now to the research study, a deep dive by myself, Dr. Tyler Page from the University of Connecticut and  Fullintel Insights Manager Katie Michel into how machines process information (You can download the full white paper for the study here). Public relations professionals need to understand how AI engines select and prioritize earned media – and this study provides a framework to optimize visibility and credibility in AI-generated search.

While there were a few nuances, we found that: 

• Because AI models use retrieval-augmented generation (RAG) to locate and synthesize information, they rank content based on semantic similarly, clarity and structure, and accessibility.
• What matters to humans also matters to machines. Best practices in public relations are still best practices – even if you’re pitching to machines.

But this led us to another, deeper question: Which content characteristics of news articles influence AI engine outputs, and how can PR professionals optimize earned media for visibility and credibility in AI-generated responses?

Here’s how we built our study to answer that question:

• We started with nearly 450 distinct prompts designed to capture how people search for health information (specifically around GLP-1 and weight-loss drugs)
• We ran those prompts across 10 personas built to reflect U.S. demographics and audience psychographics
• Responses were collected from ChatGPT, Gemini, and Google’s AI Overviews
• From those responses we extracted more than 1,300 URLs cited by AI engines that came from identifiable news outlets and coded them for variables like brand presence, outlet reach, sentiment, and topic prominence

After coding and analysis, a clear theme emerged around the type of content most often cited.

How to Get Cited in AI Engines? Focus on Well-Structured, Educational Content

Our study showed a clear preference for content that teaches, explains, and informs – not promotional, not purely news-driven, but genuinely educational.

This makes intuitive sense because AI engines synthesize answers for users, and naturally want to return quality answers. But what does that look like in practice?

Content cited by AI engines consistently demonstrated five characteristics around content structure and clarity:

1. The majority of articles cited by AI were educational 
2. Most were answer-style articles structured to directly respond to a reader’s question
3. Lean, focused articles without many external links performed better
4. Content had a PageRank score of at least five (a baseline domain authority signal that AI engines appear to use as a credibility filter) 
5. Content had high topic prominence (the core topic was typically mentioned in the headline and throughout the body text)

These findings are great news for PR practitioners, because these are things you can influence by pitching differently or shaping your content with these factors in mind.

How to Get Cited by AI Engines More Frequently?

At this point, we’ve identified which characteristics help get content cited in AI engines. But we wanted to take our research deeper than that by asking: Which characteristics help certain articles get cited more frequently? 

I’ll share five key findings:

1. As above, answer-style articles performed best: We recommend focusing on educational articles that answer a specific question 
2. Comparison-style articles can increase AI citations (many AI prompts are based on similar comparisons)
3. Human-focused visuals are important. This is also reflected in Fullintel’s Media Impact Score, which predicts how likely news coverage is to influence audiences and includes several factors, including visuals (visuals with a human element get a heavier weighting).
4. AI engines love credible expert voices: The articles that got cited more often had multiple credible voices (third-party voices like doctors or other experts)
5. High topic prominence is key: Having the topic in the headline is crucial for getting cited more often

Almost as interesting as what predicts AI citation, however, is what does NOT predict it: We found that factors like outlet reach, social shares/engagement, brand mentions in headlines, and negative sentiment weren’t key drivers of AI engine citations.  

Strategic Recommendations for PR professionals 

Based on our research, here are some best practices for PR professionals to maximize earned content in AI search:

• Focus on educational content that answers questions 
• Incorporate credible sources
• Optimize topic prominence
• Use human visuals (ideally human with brand to maximize engagement, brand recall and AI citation)
• Don’t panic over one negative article or fixate on social shares or outlet reach 

Our research study gives PR professionals a practical framework for navigating earned media in the age of AI search, with evidence-backed strategies to boost both visibility and credibility.

While this specific study examined one particular industry – the pharmaceutical/healthcare space – we strongly believe similar results can be replicated in other industries. And as we continue to conduct studies, we’re confirming these results and will track any changes as model algorithms are updated. For now, the main takeaway is to not give up on best practices in public relations, and add some new key drivers for AI citation.