A reporter used to be the main gatekeeper between your story and the public. Now there are two. The reporter still decides whether you get covered. Then ChatGPT, Gemini, Claude, or Google’s AI Overviews decide whether that coverage ever gets repeated to the person asking about you.
That second gatekeeper changes the job.
AI-first PR is earning media coverage and third-party mentions so AI assistants and answer engines cite, summarize, and recommend a brand when people ask about its category, leaders, or problems. The coverage still has to persuade humans. It also has to be findable, quotable, and recent enough for a language model to pull into an answer.
Most agencies selling public relations services haven’t caught up. They’re still optimizing for the clip report. The brands that win in 2026 are optimizing for the answer.
Why does PR matter more for AI search than SEO does?
Because AI models trust what other people say about you far more than what you say about yourself.
Muck Rack’s What Is AI Reading? study, now in its third edition, analyzed more than 25 million links cited by ChatGPT, Claude, and Gemini across 17 industries. In the May 2026 release, earned media accounted for 84% of all AI citations. Across all three editions, that figure never dropped below 82%. Paid and advertorial content? 0.3%.
Read that again if you run a marketing budget. A sponsored post barely registers. A well-placed article in a trade publication, a mention in a government report, a quote in a local news story: that’s the raw material AI answers are built from.
Journalism alone made up 27% of cited links, spread across more than 20,000 outlets. That last detail matters. AI isn’t just reading the Wall Street Journal. It’s reading niche trades, regional papers, and industry newsletters, which is where most brands can actually earn placements.
SEO still matters. Your own pages need to be crawlable and clear. But your website is the witness who’s related to the defendant. AI weighs that testimony accordingly.
The click is disappearing, and PR is filling the gap
Traditional PR leaned on a quiet assumption: coverage leads to searches, searches lead to visits, visits lead to leads. That chain is breaking in the middle.
Pew Research Center tracked the browsing behavior of 900 U.S. adults in March 2025. When a Google search showed an AI summary, users clicked a traditional result in 8% of visits. Without the summary, they clicked 15% of the time. Clicks on the source links inside the AI Overview itself? About 1%. And 26% of searches with an AI summary ended the browsing session entirely, versus 16% without one. (Google disputed the methodology, calling the query set unrepresentative. The direction of the finding lines up with what most publishers are reporting anyway.)
So the person researching your company may never land on your site. They’ll read a paragraph that an AI wrote about you, assembled from whatever coverage it found. That paragraph is your reputation now.
This is why brands are moving money. An Outcomes Rocket benchmark of 858 marketing and PR professionals, published in May 2026, found that 44.8% of organizations had increased PR spending as a direct result of AI-driven search. The budget is following the citation.
Where most PR programs are falling short
PR teams know this shift is happening. Most haven’t reorganized around it.
Muck Rack’s 2026 State of PR report surveyed 971 PR professionals. About 73% said generative engine optimization (GEO) is at least somewhat important to their strategy. Yet 29% said nobody at their organization owns GEO, and 39% don’t measure GEO success at all.
That’s the gap. Everyone agrees AI visibility matters. Almost a third haven’t assigned it to a human being.
Meanwhile, the media relations grind is getting harder. In the same survey, 71% reported low journalist response rates and 61% pointed to shrinking media lists. The pool of reporters is smaller, and the ones left are buried. Muck Rack’s 2026 State of Journalism report found that 88% of journalists immediately discard pitches that miss their beat, and nearly half say they seldom receive pitches that match what they cover.
A spray-and-pray press release strategy fails twice in this environment. Journalists ignore it, and AI models don’t cite it.
What public relations services must do differently
Here’s where the work actually changes. None of this replaces relationship-driven media relations. It changes what you’re pitching, where, and how you know it worked.
- Treat AI answers as a media outlet: Before planning a campaign, run the questions your buyers actually ask through ChatGPT, Gemini, Perplexity, and Google’s AI Mode. “Who’s the best [category] firm for [use case]?” “Is [your CEO] trustworthy?” “What happened with [your company] in 2024?” Record which brands appear, which sources get cited, and where you’re missing. That’s your target media list. If an AI keeps citing a specific trade publication for your category, that outlet just moved to the top of the pitch queue.
- Pitch outlets for citation value, not just reach: A placement in a mid-sized industry publication that AI models repeatedly cite can outperform a one-off national hit that never shows up in an answer. Muck Rack’s data shows AI citation behavior varies by platform: ChatGPT leans on Wikipedia, Gemini on Reddit, Claude on academic sources like PubMed Central. A serious PR program maps those preferences by category instead of guessing.
- Lead with original data: Journalists rank original research among the most valued parts of a pitch. AI systems love it too, because a specific statistic is easy to extract and attribute. A proprietary survey, an analysis of anonymized client data, or a quarterly index gives reporters a reason to write and gives models a fact to repeat with your name attached. One good dataset can fuel a year of coverage.
- Stay recent, on purpose: More than half of the journalism citations in Muck Rack’s study came from articles published within the past 12 months, with citation volume dropping sharply after six. Coverage decays. A brand that earned a great feature in 2023 and nothing since is slowly vanishing from AI answers. PR has to run as a steady cadence, not a launch-week spike.
- Make every placement quotable: AI models extract clean, self-contained claims. Spokespeople should give reporters sentences that survive being lifted out of context: a definition, a number, a firm position. Vague quotes about “being committed to excellence” get cut by editors and ignored by models. Train executives to speak in answers.
- Use press releases for what they still do well: Press releases aren’t dead, but their role has narrowed. Muck Rack found they appear almost exclusively in AI responses about industry trends, at 3.5 times the rate seen in “best-of” queries. A release won’t get you recommended as the top provider. It can get your data or announcement folded into a trend summary. Write releases accordingly: lead with the finding, not the CEO’s excitement.
- Fix the facts at the source: AI answers inherit errors. An outdated Wikipedia line, an old lawsuit story with no follow-up, a misattributed quote in a syndicated article: these get repeated with total confidence. Modern public relations services need a correction workflow: identify the bad source, request updates from the publisher, and earn newer, accurate coverage that outweighs it.
- Connect PR, SEO, and reputation under one owner: The 29% of organizations where no one owns GEO are the ones losing ground. Earned media, owned content, review profiles, and executive bios all feed the same AI answer. When those sit in four different departments, nobody sees the whole picture the model is seeing.
How should PR success be measured in an AI-first world?
Impressions and ad-value equivalency were weak metrics before AI. Now they’re close to meaningless, because the reader may never see the original article. They see the summary.
Better measures look at what the AI actually says:
- Share of voice in AI answers: How often does your brand appear when a model is asked a category question, compared with competitors? Track it weekly across platforms, using a fixed set of prompts.
- Citation sources: Which of your placements are getting cited? This tells you which outlets are worth the next pitch.
- Sentiment and accuracy of AI descriptions: Is the model describing you correctly? Is it surfacing a ten-year-old controversy as current?
- Branded search and direct traffic trends: People who see you named in an AI answer often search for you directly afterward. That lift is one of the few click-based signals still worth watching.
Tools now exist for most of this. Ahrefs Brand Radar, Muck Rack’s AI visibility features, and similar platforms let teams track mentions and citations across AI assistants over time. NetReputation, for example, monitors how often AI assistants name it in response to “best reputation management company” prompts, and watched its ChatGPT presence and Gemini presence move in different directions over the same period. That kind of platform-by-platform view is the only way to know where a campaign is working.
The uncomfortable part: you can’t fully control any of this. Models change their sourcing behavior without notice. What you can control is the volume, quality, and freshness of credible third-party coverage about you. That’s still PR’s job. The scoreboard just moved.
Crisis communications runs on a shorter fuse now
A negative story used to have a news cycle. It peaked, it faded, and eventually it slipped to page two of search results.
AI compresses that. A model asked about your company today might summarize last month’s incident in its first sentence, with no context and no mention of how you responded. Because recent journalism gets cited most heavily, a fresh negative story can dominate AI answers precisely when you need them to be balanced.
That changes crisis playbooks in a few concrete ways. Your response statement should be written to be quoted, because the model may only pull one line. Follow-up coverage of corrective action matters more than ever, since it’s the only thing that can sit alongside the original story in an answer. And monitoring has to include AI outputs, not just news alerts and social listening. If you only watch Google News, you won’t see what ChatGPT is telling your customers.
Firms that combine crisis management with reputation and AI search work, like NetReputation, approach this as one problem: the incident, the coverage, and the AI summary of both.
What to ask before hiring a PR firm in 2026
If you’re evaluating public relations services right now, the pitch deck will look familiar. Ask questions that reveal whether the firm has actually adapted:
- How do you decide which outlets to target? (Listen for AI citation data, not just circulation numbers.)
- What do you report on monthly besides placements? (You want AI share of voice and citation tracking.)
- How do you handle inaccurate information that AI models keep repeating?
- What original research have you helped clients produce, and where did it get cited?
- Who on your team owns GEO, by name?
A firm that answers the last question with a long pause has told you what you need to know.
The bottom line for PR teams
The fundamentals of public relations haven’t changed. You still need a real story, a credible spokesperson, and a reporter who trusts you. What changed is who’s reading the coverage. Half the audience is now a machine deciding what to tell everyone else.
PR teams that keep chasing impressions will produce nice clip reports and lose the answer. The ones that pitch for citation value, publish original data, stay recent, and measure what AI actually says will own their category’s narrative, whether the buyer ever clicks or not.
Frequently Asked Questions
What are public relations services?
Public relations services help organizations earn credible third-party attention through media relations, thought leadership, crisis communications, and reputation strategy. In 2026, strong PR services also track how AI assistants describe a brand, because earned coverage is now the main source material ChatGPT, Gemini, and Google’s AI Overviews use to answer questions about companies.
How does AI change public relations?
AI assistants now summarize brands directly, so many people form opinions without visiting a website. Since earned media makes up about 84% of AI citations in Muck Rack’s research, PR coverage shapes what those summaries say. PR teams must pitch outlets AI cites, publish quotable data, and keep coverage recent.
What is generative engine optimization in PR?
Generative engine optimization (GEO) in PR means earning and structuring coverage so AI models cite and recommend your brand. It includes targeting outlets that AI systems frequently reference, providing original data reporters can attribute, keeping coverage fresh, and monitoring AI answers for brand share of voice, accuracy, and sentiment over time.
Do press releases still help with AI visibility?
Yes, within limits. Muck Rack found press releases appear mostly in AI answers about industry trends, at 3.5 times the rate seen in “best-of” recommendation queries. A release can get your data into trend summaries, but earned editorial coverage is what gets a brand recommended as a top provider.
How do you measure PR results when AI reduces clicks?
Track share of voice across AI assistants using a fixed set of category prompts, identify which placements AI cites, and review the accuracy and sentiment of AI descriptions of your brand. Pair those with branded search and direct traffic trends, which often rise after a brand is named in AI answers.
View 0 comments