The question comes up constantly among communications teams evaluating where to spend a limited budget: does sending a press release over the wire actually help a company show up when someone asks ChatGPT or Perplexity a relevant question? For a while the honest answer was “probably, a little, we think.” That’s changed. Several research groups have now published large-scale audits of what AI models actually cite, and the picture is much clearer than it used to be.
This article walks through the current data, where it comes from, and what it means for a company deciding whether press release distribution is worth the spend.
The Short Answer From Recent AI Citation Research
Across every major study published so far in 2026, the answer lands in roughly the same place: press releases contribute a small, measurable amount to AI visibility, and earned media contributes the overwhelming majority.
Muck Rack’s Generative Pulse research, summarized in detail by AuthorityTech’s breakdown of the study, analyzed more than 25 million links cited by ChatGPT, Claude, and Gemini across 17 industries. Press releases accounted for less than 2 percent of those citations, while earned media accounted for the large majority.
Meltwater’s April 2026 GenAI Lens analysis, which tracked roughly 5.35 million citations across eight major AI platforms including ChatGPT, Claude, Gemini, and Perplexity, found press release citations fell within a single month during the period studied, even as citations from earned and news media grew over the same window.
Where The Numbers Come From
It’s worth understanding how these figures were produced, since the methodology affects how much weight to put on them.
Large-Scale Citation Audits Across Multiple Models
The strongest studies in this area don’t rely on a handful of test queries. Muck Rack’s research spans millions of citations across three major models and dozens of industries, and Meltwater’s tracking pulls from eight platforms simultaneously, including ChatGPT, Grok, Claude, Perplexity, Gemini, and both Google AI Overviews and Google AI Mode. That breadth is what makes the consistency across studies meaningful rather than coincidental.
Independent Academic Research Points The Same Direction
Academic researchers have reached similar conclusions from a different angle. A study examining how language models select sources found that outlet authority, not content quality alone, is the primary driver of which sources get cited, a pattern consistent with why individually distributed press releases struggle to compete with established news outlets for citation share.
Why The Numbers Look This Way
The data isn’t just a curiosity. It reflects specific, explainable behavior in how AI retrieval systems process content.
Duplicate Wire Content Gets Filtered
Presenc AI’s research on distribution and AI visibility found that raw syndicated releases, meaning identical text appearing across hundreds of newswire pickup sites, are largely deduplicated by AI retrieval systems and rarely cited directly. A release that generates 200 identical pickup pages and zero original articles was found to create minimal AI citation value.
Earned Coverage Signals Independence
The same Presenc AI research found that releases generating genuine editorial coverage, meaning articles written independently by journalists at recognized outlets, created an estimated three to five new citation sources in AI answer pipelines within 30 to 60 days. The difference isn’t the underlying news. It’s whether an independent third party chose to write about it.
What The Data Means For Companies Weighing Distribution
None of this data argues that press releases are worthless, but it does argue for a specific, narrower expectation of what they accomplish.
Set Realistic Goals For What A Release Alone Will Do
Based on the current research, a company shouldn’t expect a distributed release to be cited by name in AI answers with any regularity. The data consistently puts that outcome in the low single digits at best.
Weigh Distribution Against Its Actual Function
Additional analysis tracking domain authority and AI citation correlation found that a single placement in a top-tier outlet can generate more AI visibility than dozens of smaller mentions combined, and that branded web mentions correlate far more strongly with AI visibility than backlinks do. That points toward evaluating distribution by the original coverage and brand mentions it generates, rather than by its own direct citation rate.
Frequently Asked Questions
What percentage of AI citations come from press releases?
Recent research from Muck Rack found press releases account for less than 2 percent of citations across ChatGPT, Claude, and Gemini.
Meltwater’s April 2026 tracking found press release citations declined within a single month studied, while earned media citations grew over the same period.
Why do AI models rarely cite press releases directly?
Identical release text syndicated across many wire pickup sites tends to get deduplicated by AI retrieval systems, which reduces its citation value regardless of distribution volume.
Does this data mean press releases have no value for AI visibility?
Not entirely. Releases that generate genuine, independent editorial coverage can indirectly produce new AI citation sources, even though the release itself is rarely the thing cited.
The data on this question has become fairly consistent across independent studies, which makes it easier to plan around than it was even a year ago. Press releases alone rarely earn AI citations. What they can still do, when written and distributed well, is generate the kind of original coverage that does.