82% of Journalists Now Research Stories in AI. Here's What Happens When Your Brand Isn't in the Answer.
Muck Rack's 2026 State of Journalism report shows 82% of journalists use AI tools for research. If your brand is absent from AI answers, you're invisible to both machines and the editors who now rely on them. Here's the three-step audit to close the loop.
There is a feedback loop forming in earned media that most marketing and communications teams have not mapped yet. Journalists are researching stories inside AI tools. AI tools are citing earned media. And the brands that earned media covers are the ones AI tools surface in the first place.
If your brand is not in AI answers, journalists using those answers to source stories are less likely to find you. If journalists are less likely to find you, you earn fewer placements. Fewer placements mean less material for AI engines to cite. The loop tightens, and you fall further behind with each cycle.
Muck Rack's 2026 State of Journalism report puts the number at 82%. ChatGPT leads at 47% adoption among working journalists, with Google Gemini at 22% and climbing (Muck Rack, 2026). These are not casual users. They are reporters using AI to identify sources, surface background research, and find companies relevant to a beat they are covering right now.
That changes the dynamics of media targeting in a way most PR teams have not absorbed.
The 2% overlap problem
The signal that makes this concrete comes from Muck Rack's Generative Pulse analysis of more than one million AI-generated citations across ChatGPT, Claude, Gemini, and Perplexity. The finding: the journalists most frequently pitched by PR professionals and the journalists most frequently cited by AI engines overlap by just 2% (Muck Rack Generative Pulse, via GlobeNewswire, March 30, 2026).
That is not a rounding error. It is a structural mismatch between where PR effort goes and where AI citation value concentrates.
Christian Lehman's read on this: the 2% figure exposes a broken assumption baked into most media targeting workflows. Teams build pitch lists around who covers their beat, who has the biggest audience, and who responded last quarter. None of those criteria map to which journalists produce content that AI engines retrieve and cite. The overlap gap means most PR budgets are optimizing for human reach while ignoring the machine-readable authority signal that earned media is uniquely positioned to produce.
The same Generative Pulse data shows earned media accounts for nearly 25% of all LLM citations, with non-paid sources representing 94% of all AI-cited links (Muck Rack Generative Pulse, March 2026). Ahrefs' 75,000-brand study confirmed that brand web mentions correlate 3x more strongly with AI visibility than backlinks (0.664 vs. 0.218) (Ahrefs, 2025). The Fullintel-UConn academic study presented at IPRRC in February 2026 found 47% of all AI citations came from journalistic sources, with 95% of cited links being unpaid (Fullintel, 2026).
The earned media pipeline is the single largest input to what AI engines recommend. And 82% of the journalists who produce that pipeline are now pre-filtering their source lists through the same AI engines.
How the recursive loop works
This is not a metaphor. It is a measurable feedback cycle with four stages.
| Stage | What happens | Who is affected |
|---|---|---|
| 1. Journalist research | Reporter opens ChatGPT or Perplexity to find companies and sources for a story | Brands absent from AI answers are structurally excluded from consideration |
| 2. Story publishes | Article names (or omits) your brand based on what the journalist found in AI research | Omitted brands lose a placement opportunity they never saw |
| 3. AI crawlers index | AI engines index the new article within hours, incorporating brand mentions into the citation graph | Brands mentioned gain citation compounding; brands omitted gain nothing |
| 4. Next journalist searches | The next reporter researching the same category gets an updated answer that reflects the new article | The loop repeats, wider for brands inside it, narrower for brands outside |
A journalist at a trade publication opens ChatGPT to research which companies are leading a particular category. The AI assembles its answer from the publication graph it trusts, pulling names that appear across multiple third-party sources. If your brand has earned placements in those publications, you show up in the journalist's research. If you have not, you are structurally absent from the moment the story starts forming.
This is where the compounding data becomes urgent. The Authoritas study tracking 143 digital marketing experts found that between December 2025 and February 2026, the top 10 captured 59.5% of all citability across ChatGPT, Gemini, and Perplexity, up from 30.9% two months prior. The Herfindahl-Hirschman Index of citation concentration rose 293% in under two months (Authoritas, 2026).
The window for entering the loop is narrowing. Brands already inside are compounding. Brands outside are watching the gap accelerate.
The three-step media targeting audit
If your team is running earned media and has not mapped targeting to AI citation behavior, Christian Lehman recommends starting with this audit. It takes one afternoon and tells you whether your pitch list is aligned with the publications AI engines actually cite.
Step 1: Map your current pitch list against AI citation sources. Pull your team's active media list. For the top 20 targets, search each publication name in the Muck Rack Generative Pulse data or run each through a direct AI query. Ask ChatGPT or Perplexity: "Which [publication name] articles are most frequently referenced when answering questions about [your category]?" If fewer than five of your top 20 targets appear in AI citation results for your category, your list has the 2% problem.
Step 2: Identify the publications AI engines actually cite for your category. Run five category queries across ChatGPT, Google AI Mode, and Perplexity. Record every publication cited in each response. Build a frequency-ranked list. The publications that appear most often across engines and queries are the ones producing earned authority that compounds. Cross-reference that list against your pitch targets. Any high-citation publication not on your list is a gap.
AT's research on earned media vs. owned content citation rates found that earned media distributed through trusted publications generates 325% more AI citations than brand-owned content. That gap means targeting the right publications is not a marginal optimization. It is the difference between building citation infrastructure and producing content nobody cites.
Step 3: Prioritize editors whose coverage produces AI-cited content. Within the high-citation publications from Step 2, identify the specific journalists whose articles appear in AI answers. This is the reverse of the traditional media list build. Instead of starting with beat coverage and working outward, you start with which reporters produce content that AI engines retrieve and cite, then work backward to relationship-building.
The Muck Rack 2026 State of Journalism data gives additional context for what these journalists respond to: 88% immediately delete pitches not aligned with their beat, 70% prioritize beat alignment, 58% want access to credible sources, and 40% value original data (Muck Rack, 2026). Pitches under 200 words are preferred by 69% of respondents.
| What AI-cited journalists want | What most PR teams send |
|---|---|
| Beat-aligned, data-rich pitches under 200 words | Generic category overviews sent to broad lists |
| Access to credible, named sources with original data | Executive quotes without supporting numbers |
| Specific findings tied to trends they are already covering | Product announcements framed as industry news |
That profile maps directly to the kind of content AI engines also prefer to cite: data-rich, specific, attributed to named sources. The pitch that lands with the AI-cited journalist and the content that earns AI citations share the same structural DNA.
Why this changes what earned media produces
The implication is not that PR teams should abandon human-audience targeting. It is that the audience for a Tier 1 placement is now two audiences simultaneously. The human reader who makes a decision based on the article. And the AI system that indexes the article within hours and starts using it to construct answers for thousands of future queries.
A placement in a publication that AI engines heavily cite is doing double duty. It builds human credibility and machine-readable authority at the same time. A placement in a publication that no AI engine retrieves produces human reach and zero citation compounding.
The Pew Research Center's July 2025 study found that click rates halve when AI summaries appear in search results, dropping from 15% to 8% (Pew, 2025). The Bain 2026 consumer study showed 80% of search users rely on AI summaries at least 40% of the time (Bain, 2025). Human discovery is being mediated by AI at increasing rates. The earned placements that feed both channels are the highest-leverage investments a growth team can make.
As Christian Lehman has written previously, the execution challenge is not producing more content but targeting the right publications. The three-step audit above is where that targeting rebuild starts.
This is where Machine Relations provides the operating frame for what is otherwise just a measurement insight. The mechanism is simple: earned media in publications AI engines trust creates third-party credibility that machines can index, retrieve, and cite. That citation makes the brand appear in AI answers. Those AI answers are now the first research step for both buyers and the journalists who write the next round of coverage.
The loop is live. The question is whether your brand is inside it or watching it run without you.
Run the three-step audit this week. If your media list has the 2% problem, the fix is not more pitches to the same targets. It is a targeting rebuild around the publications and journalists whose work AI engines actually use. AuthorityTech's visibility audit maps your current earned media footprint against the publications AI engines cite for your category, so you know exactly where the gap is before you brief a single pitch.
Key takeaways
- 82% of journalists now use AI tools for story research, with ChatGPT at 47% adoption and Gemini at 22% (Muck Rack, 2026)
- The journalists most pitched by PR teams and those most cited by AI engines overlap by only 2% (Muck Rack Generative Pulse, 2026)
- Earned media accounts for 25% of all LLM citations; 94% of AI-cited links are non-paid (Muck Rack, 2026)
- Citation concentration among top entities rose 293% in two months, and the compounding is accelerating (Authoritas, 2026)
- Brand web mentions correlate 3x more strongly with AI visibility than backlinks (Ahrefs, 2025)
FAQ
Why does the 2% journalist overlap matter for brand visibility? The 2% overlap between most-pitched journalists and most-AI-cited journalists means most PR spend is targeting reporters whose work does not appear in AI answers. Since 82% of journalists now use AI for research, brands absent from AI answers are also absent from the sourcing pool reporters draw from. The result is a compounding exclusion from both AI citation and future earned coverage.
How do I find which publications AI engines cite for my category? Run five category queries across ChatGPT, Google AI Mode, and Perplexity. Record every publication cited. Build a frequency-ranked list. The publications that appear most often are the ones producing citation infrastructure for your space. Machine Relations, coined by Jaxon Parrott in 2024, defines this as the earned authority layer that makes brands the answer AI engines give.
What makes a press release earn AI citations? Muck Rack's Generative Pulse data found cited press releases contain roughly twice as many statistics, 30% more action verbs, 2.5 times as many bullet points, and a 30% higher rate of objective sentences compared to uncited releases. AI engines reward data-rich, specific, substantively written releases, not boilerplate.