- Journalism alone accounts for 25-27% of all AI citations across three independent Muck Rack "Generative Pulse" studies since July 2025, a large, stable share of the earned-media citation pool covered in why earned media outperforms owned content.
- Earning that coverage is a supply problem before it's an outreach problem: Cision's 2023 State of the Media Report (3,132 journalists, 17 markets) found journalists roughly 78% more likely to cover a story offering exclusive data unavailable to competing outlets, original research doesn't just add color, it can be the reason a story runs at all.
- That figure is a few years old now. It's cited here as the most specific disclosed number found on this exact question, not as a freshly re-confirmed one.
- Press releases and paid/advertorial content sit at the bottom of what actually earns a citation: 0.04% and 0.3% of AI citations respectively, per two independent datasets (BuzzStream/Citation Labs and Muck Rack) that land almost exactly on the same near-zero number.
- This piece is the "what to have ready" layer. The mechanics of actually reaching a journalist (format, timing, response rates) are covered separately in how to pitch a journalist for a citation-worthy mention.
Why earned media outperforms owned content in AI citations already establishes that journalism alone accounts for 25-27% of all AI citations, consistently, across three independent Muck Rack studies since July 2025. What that piece doesn't cover is how a specific piece of coverage actually gets earned. The honest answer is that it's mostly not a pitching problem. It's a "do you actually have something a journalist would want to cite" problem, solved before the first email ever goes out.
What do journalists say they actually want?
Cision's 2023 State of the Media Report, 3,132 journalists surveyed across 17 markets (US, UK, Canada, and others), fielded February-March 2023, found journalists roughly 78% more likely to cover a story that offers exclusive data unavailable to competing outlets. That's a specific, disclosed-sample figure, though a few years old now; it's cited here because it's the most concrete number found on this exact question, not because it's been freshly re-measured in a more recent edition of the same survey.
The pattern across all of this is the same: a journalist's job is to publish something readers haven't already seen elsewhere. A press release restates something the brand already wants said. An expert quote adds color to a story that already exists. Original data is the one input that can be the story, which is why it consistently outranks the rest of what PR teams typically lead with.
Why does this specifically favor an unknown brand?
This matters more for a brand nobody's heard of than for an established one. An unfamiliar founder or company has no existing relationship, no reputation a journalist can lean on to justify the coverage, and no obvious reason to get a callback over the dozens of other pitches competing for the same inbox. Original data sidesteps all of that: a journalist doesn't need to already trust or know the source to use a specific, checkable number nobody else has published. The data does the credibility work that a relationship would otherwise have to do.
What consistently doesn't work?
Two independent datasets converge on the same finding from different directions. BuzzStream/Citation Labs' analysis of 30,000 citations across 595 prompts and 10 industries found press releases earn roughly 0.04% of AI citations. Muck Rack's separate Generative Pulse data puts paid and advertorial content at 0.3%. Different organizations, different samples, both landing near zero: a press release announces something happened; it doesn't hand a journalist independent evidence to build a story around, which is the actual threshold that gets material covered and, downstream, cited.
What does "citation-worthy" material actually look like?
Based on what the surveyed preference data above actually rewards, the material worth having ready before any outreach starts falls into a short list:
- A number nobody else has published. Your own usage data, a survey you ran, a benchmark you built: something that didn't exist publicly before you published it.
- A finding that updates or complicates an existing narrative, not one that simply confirms what's already assumed. A journalist covering AI search adoption has read a dozen surveys saying usage is rising; a specific, disclosed-methodology finding that complicates or sharpens that story is more valuable than the thirteenth confirmation of the trend.
- A disclosed methodology, even a modest one: sample size, date range, how it was collected. Where "X% of people use AI search" stats actually come from documents exactly what happens to a number that circulates without one: it gets generalized past what it actually measured, and eventually a piece like that one has to go trace it back.
- A live, willing, responsive human attached to the data: a named source willing to talk on a short deadline, since journalism runs on tighter timelines than most B2B content cycles.
Why is this specifically underexploited for AI-visibility topics right now?
Muck Rack's 2026 State of Journalism survey (897 journalists, fielded January-March 2026) found 82% of journalists already use AI tools themselves: meaning a meaningful share are actively reporting on how AI products and AI search actually behave, not just using AI as a writing aid. That's a narrower, currently less-saturated beat than general business or tech PR: original data about AI search behavior, AI citation patterns, or AI visibility specifically has a real chance of being the exact kind of exclusive material a journalist covering that beat is short on, compared to pitching into an oversaturated general-tech news cycle.
Frequently asked questions
What's the single most reliable way to earn journalist coverage?
Original data a journalist can't get anywhere else. Cision's 2023 State of the Media Report (3,132 journalists surveyed across 17 markets) found journalists roughly 78% more likely to cover a story offering exclusive data unavailable to competing outlets. That figure is a few years old now. This piece cites it because it's the most specific disclosed number available on this exact question, not because it's been re-confirmed in a newer edition.
Do journalists actually cite AI-visibility content specifically, or just general PR?
The same mechanics apply, with one added advantage: AI-visibility and search-behavior data is a topic journalists covering the AI beat are actively looking for sources on right now, which is a narrower, currently-underserved niche compared to a saturated general-business pitch. Muck Rack's 2026 survey (897 journalists) found 82% already use AI tools themselves, and a meaningful share cover AI as a beat directly: original data about how AI search or AI visibility actually behaves is exactly the kind of exclusive, timely material this piece describes.
Is a press release enough to earn a citation?
No, and the data on this is unusually consistent. An independent 30,000-citation dataset from BuzzStream/Citation Labs found press releases distributed via syndication channels (Yahoo, MSN, and similar) accounted for only about 0.04% of the entire dataset, landing almost exactly where Muck Rack's separate 0.3% figure for paid/advertorial content sits. A press release announces something; it doesn't hand a journalist evidence they can build an independent story around, which is the actual bar.
How is this different from just pitching a journalist?
This piece covers what has to be true before a pitch is worth sending: the underlying material a journalist would actually want to cite. How to pitch a journalist for a citation-worthy mention covers the execution once that material exists: format, timing, response-rate data, and where to find the requests in the first place.
Sources referenced
- Muck Rack Generative Pulse, "What Is AI Reading?" (Jul 2025 / Dec 2025 / May 2026 editions, 25M+ links)
- Muck Rack: 2026 State of Journalism Report (Jan-Mar 2026, n=1,044 surveyed / 897 analyzed)
- Cision: The 2023 Global State of the Media Report (Feb-Mar 2023, n=3,132 journalists, 17 markets)
This piece is the "what to have ready" companion to how to pitch a journalist for a citation-worthy mention. For the underlying data on why earned media dominates AI citation share in the first place, see why earned media outperforms owned content in AI citations.
