Generative Pulse: earned media 84% citations

- Muck Rack’s Generative Pulse said on May 7 that earned media supplied 84% of citations in AI answers across ChatGPT, Claude, and Gemini. - The standout detail is what barely shows up: paid and advertorial material made up just 0.3% of cited links, while journalism alone supplied 27%. - That matters because GEO is shifting marketing from ranking pages to winning citations inside AI answers.

Generative search is turning PR into infrastructure. That is the real story here. Muck Rack’s Generative Pulse said this week that 84% of the links cited in AI answers come from earned media, not paid placements, and that pattern has held across three reports since July 2025. (markets.businessinsider.com) ### What counts as “earned media” here? Basically, it is everything a brand does not directly buy as an ad slot — journalism, academic research, government pages, encyclopedic references, reviews, forums, and third-party corporate mentions. In the May 2026 readout, that whole bucket made up 84% of citations in answers generated by major AI systems, while paid and advertorial content was just 0.3%. (manilatimes.net) ### Why is 84% such a big deal? Because it tells marketers that AI systems are not behaving like display ad networks. They are behaving more like citation engines. If a model needs support for a claim, it reaches for material that looks edit(manilatimes.net 1)(manilatimes.net 2) ### Where does journalism fit? Journalism is still a huge piece of the stack. The May report says news content alone accounted for 27% of all cited links. That lines up with Muck Rack’s earlier launch research, which also put journalistic co(manilatimes.net)rage when the question is fresh. (manilatimes.net) ### Does this mean SEO matters less? Not exactly — but the unit that matters is changing. Search Engine Land’s GEO framework says brands now need to track things like AI citation frequency, share of voice in generative results, referral traf(manilatimes.net)t position is my page in?” (searchengineland.com) ### Why would AI prefer earned sources? Turns out earned sources usually carry the signals models need most — independent wording, clearer attribution, and repeated mention across multiple places. An ad page can say anything about itself. A reported article, government page, or widely referenced explainer looks more like evidenc(searchengineland.com)sier for retrieval systems to trust. This is an inference from the citation mix and from how GEO measurement is being framed. (searchengineland.com) ### What changes for brands? The playbook shifts upstream. Instead of only optimizing owned pages for keywords and backlinks, brands now need coverage, expert mentions, quotable research, and pages that can survive being excerpted by an AI. If nobody independent is talking about you, the model has less reason to mention you. And if it does mention you, it may pull framing from someone else. (markets.businessinsider.com) ### What is the catch? Being cited is not the same as getting traffic. Google is also adding more direct source links and richer source previews inside AI Overviews and AI Mode, which suggests the platforms know source visibility is a pressure point. But publisher and brand sites still face a world where the answer often arrives before the click. (9to5google.com) ### Bottom line? The old search fight was over ranking. The new one is over being the source the model trusts enough to quote. In that world, PR, publisher coverage, and third-party validation stop being support functions and start looking like core search infrastructure.

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.