web analytics

Strategy

November 18, 2026
Share To:

The Data Behind AI Visibility

By Ryan Heath

Why your experts are your most underrated AI search asset.

Search has flipped. People used to type fragments—“best faucet finish for hard water”—then wade through 10 blue links. Now, they ask the full question and expect one answer back. And that answer is only as good as the sources the AI chose to trust. If your brand isn’t in that source list, you’re not ranking lower. You’re not even in the room.

Public relations has always been about shaping how people see a brand. That part hasn’t changed. What’s changed is the audience—it now includes the AI models fielding your customers’ questions every day. And a run of recent research points pretty clearly to where the opportunity is, much of it playing to the strengths of a good PR and content program. (For the case on why PR belongs at the center of all this, our colleague Rosie made that argument earlier this year.)

What is AI actually reading?

Muck Rack’s “What Is AI Reading?” study from May 2026 gives us the best look yet at where large language models pull their answers from. For anyone investing in earned media, the numbers are encouraging:

  • 84% of all AI citations come from earned media—journalism, research, user-generated content, and third-party content.
  • 27% come from journalistic sources specifically—the biggest single category.
  • Paid and advertorial content? Under 1%. In an AI-mediated world, you get cited by earning it.


Press releases are a quiet winner here, especially when they’re tied to a genuine industry trend. Trend-anchored releases show up in AI answers 3.5x more often for trend queries than for “best-of” queries, and trend questions pull journalism citations at twice the rate of how-to questions. A well-timed release remains one of the few brand-controlled assets AI actively surfaces, which makes strategic PR more valuable now—not less.

LinkedIn is the sleeper engine of AI visibility.

While Muck Rack’s study tells us what kind of content AI trusts, Meltwater and LinkedIn’s report on how LinkedIn content wins in AI search—drawn from 9.5 million citations—tells us where a lot of it actually lives:

  • LinkedIn is the #2 most-cited domain across every major AI model in B2B categories, trailing only YouTube.
  • It lands in the top five in 14 of 16 B2B categories, and takes the #1 spot in Marketing & Advertising and in AI & Data Science (even topping YouTube in those two categories).
  • It beats Reddit, Medium, Quora, Wikipedia, and the major review platforms on B2B queries.


The trend is only picking up speed. An OtterlyAI analysis from June 2026 found LinkedIn’s monthly AI citations rose roughly 50% in five months. For getting a brand’s expertise in front of AI, few channels look more dependable right now.

Why your experts beat your brand page.

This is the finding that should change how teams think about content: AI reaches for people far more often than pages.

  • 75% of LinkedIn citations come from individual profiles, not Company Pages—and that held true across every model tested.
  • C-suite voices earn the highest citation rates per article.
  • Expertise beats reach. More than half of cited LinkedIn members have fewer than 10,000 followers. Audience size isn’t the currency here. Demonstrated authority on a specific subject is.


So, if your strategy funnels every insight through the Company Page and treats employee posts as optional amplification, you’re leaving your strongest asset on the bench. A VP of engineering writing plainly about material fatigue thresholds may carry more weight with AI than a polished corporate announcement because it reads like a real expert answering a real question.

What content actually gets cited.

The research lands in the same place on format. AI leans toward structured, decision-support content:

  • Top-cited formats: “Best X” listicles (54% of top-cited LinkedIn URLs), side-by-side comparisons (50%), and “how to choose” guides (33%).
  • Structure isn’t optional: Every top-cited article used bullet lists, and 92% used clear headings.
  • Length has a sweet spot: 1,500 to 2,500 words, with real names and real numbers throughout.


If a piece reads like this one—headings, bullets, named sources, hard data—it’s built the way AI wants to read it.

Recency is a habit, not a one-time push.

The last lever in the Muck Rack data is time. Content ages out, and fast:

  • 57% of journalism citations were published in the past 12 months.
  • Citation rates spike in month one and fall off sharply by month six.
  • 48% of cited LinkedIn content went up in the last three months.


Publish once and then go dark, and you hand the ground to Reddit threads, Quora answers, and content nobody at your company vetted. Staying visible isn’t a campaign. It’s a rhythm.

The playbook.

  1. Put your experts out front. Get your subject-matter experts and leadership publishing original, well-structured content on LinkedIn—regularly, not once a quarter.
  2. Write for decisions. Comparisons and “how to choose” guides, not just announcements.
  3. Tie releases to trends. That’s where AI is most likely to find them.
  4. Format for the machine as well as the reader. Headings, bullets, named entities, and that 1,500-to-2,500-word range.
  5. Make freshness a metric. Keep a refresh calendar so your best work doesn’t quietly age out at month six.


Let’s build your visibility engine.

In an AI-mediated world, earned media is trust currency, and the latest research shows us exactly where to spend it. It’s your experts, your trend-timed PR, and your steady stream of fresh content that AI is listening for. If you want to turn your team’s expertise into a measurable AI-visibility asset, let’s talk.

Related Posts

November 16, 2026
Every brand team has the same “successful social media presence” objectives: post consistently, incorporate lots of video, answer comments, track
Strategy
November 9, 2026
The buyer journey has gone dark, and the old ways of proving what worked no longer see most of what’s
Strategy
October 26, 2026
Audiences are starting to trust content that feels genuinely human over content that simply looks expensive.
Strategy