Our research

How does republished PR (churnalism) seed what AI says about a company?

By Pressfront Research · our churnalism study, read against Muck Rack and Ahrefs

Last updated: September 14, 2026

Through repetition, not through the link. A single press release, reprinted with little editing across dozens of outlets (the practice called churnalism), turns one statement into many pages that each place your name next to your category. Models learn and retrieve from text, so those repeated co-occurring mentions are exactly the signal they ingest. That is why wire distribution can lift AI visibility: not because the syndicated link carries authority (it usually carries none, being nofollow and self-published) but because the same name-near-topic pattern now sits in many places. The honest limit: this builds background familiarity and does close to nothing for competitive "best X" shortlist queries, where curated review and comparison content wins instead.
~80%of UK quality-press stories not wholly original; ~54% traceable at least partly to PR (Cardiff study)
84%of links AI models cite are earned media, not paid or advertorial (Muck Rack)
~0wire releases surfacing in our open "best X" shortlist runs

The pathway, step by step

Start with what churnalism actually is. Understaffed newsrooms reprint ready-made release copy with light editing, so one release becomes many near-identical articles across many domains. The Cardiff University study behind Flat Earth News found roughly 80% of UK quality-press stories were not wholly original and about 54% carried at least some PR-sourced material. That is the supply side of the machine.

Now look at what that machine hands an AI. It is not one authoritative citation. It is dozens of pages that each say your name next to your topic. Language models do not read link graphs the way old search did. They learn statistical associations from text and, when grounding, retrieve passages of text. Both of those care about how often and how consistently your name appears near a subject. Churnalism is a co-occurrence multiplier. It takes one mention and stamps it across the corpus.

This is the part most PR pitches get exactly backwards. They sell you the backlink. But a syndicated wire link is typically nofollow and lives on a page the release itself planted, so as a ranking or authority signal it is near worthless, which is our long-standing position. The value was never the link. It is the repeated co-occurrence the syndication leaves behind.

The honest limit, said plainly

Here is where most vendors stop talking, so we won't. Building co-occurrence through wire distribution moves the background: whether a model has any sense that you exist and what you do. It does not win the sharp competitive question. In our own recommendation study, when we asked engines open "best X for a small business" questions, the names that surfaced were the heavily-reviewed category leaders (Shopify 12 times, Wix 8, Squarespace 7), not companies whose footprint is a pile of reprinted releases. Wire releases were effectively absent from those shortlists. So if your goal is to be the answer to "best CRM," a press release is not the lever. If your goal is that the model knows you exist and describes you correctly when asked directly, co-occurrence built through real coverage is doing quiet, real work.

How we got here

Our churnalism work: the mechanism is that one release republished across many outlets creates many co-occurring mentions of a name near its topic, which is the unit models ingest. Grounded in the Cardiff/Lewis study (~2,000 UK stories, ~80% not wholly original, ~54% partly PR-sourced) and the newsroom-to-PR headcount gap documented by Pew.

Muck Rack, "What Is AI Reading?" (25M+ cited links, ChatGPT/Claude/Gemini, 17 industries): 84% of the links these models cite are earned media, versus about 0.3% paid or advertorial. Editorial pages, the kind reprinted release copy lands on, are overwhelmingly what AI pulls from. (Muck Rack sells PR software, so read the framing with that in mind; the composition finding still holds.)

Ahrefs, December 2025, 75,000 brands: branded web mentions are the strongest non-video predictor of AI visibility, at a Spearman correlation of 0.664 on ChatGPT, well above raw backlinks. Mentions, not links.

Put together: Muck Rack says AI reads earned editorial. Ahrefs says mentions predict citation better than links. Our mechanism says churnalism manufactures earned-editorial mentions at scale from a single source. Chain them and you get a claim none states alone: a syndicated release helps AI visibility by mass-producing the exact input (co-occurring editorial mentions) that both the composition data and the correlation data point to, while the link it carries, the thing it is usually sold on, is the part that does the least. And our recommendation data draws the boundary: this builds presence, not shortlist wins.

One caveat, once: our churnalism finding is mechanism plus a small recommendation sample, and the Muck Rack and Ahrefs figures are single-vendor studies with a product to sell. Directional, not definitive. The direction, that PR works for AI through repeated mentions rather than through the backlink, is the part we would defend.

What this means if you distribute PR

Distribute for the reprint footprint, not the link. Wider syndication onto real editorial domains is more co-occurring mentions of your name near your category, which is the thing that actually registers. Just do not confuse that with winning "best X," and do not let anyone sell you the nofollow link as if it were the point. It never was.

Want to see whether your coverage has actually put you in the models' picture of your category? Pressfront's free AI Visibility Report checks whether the real engines know you exist and how they describe you.

Related entries

Changelog

September 14, 2026: First published. Converts the churnalism study into the syndication-to-citation pathway and crosses it with Muck Rack's earned-media composition and Ahrefs' mentions correlation.

Sources (primary): Pressfront Research churnalism analysis (mechanism), drawing on the Cardiff University study led by Justin Lewis reported in Nick Davies, Flat Earth News (2008), overview at Churnalism, and Pew Research Center, "The growing pay gap between journalism and public relations" (2014); Muck Rack, "What Is AI Reading?" (25M+ links, ChatGPT/Claude/Gemini), 84% earned media; Ahrefs, "Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews", 75,000 brands, December 12, 2025; Pressfront Research, Run-to-Run Consistency of Business Recommendations, 60 responses, DOI 10.5281/zenodo.22738861.