A comms leader I work with lost a third of her team about two years ago.

The CMO made a compelling case: Advertising had measurable return on ad spend (ROAS). Content marketing had attribution, earned media had neither. It wasn’t an unreasonable argument. The metrics said what they said. Her headcount went to marketing.

We ran an AI visibility audit earlier this month for them. The brand was barely present. Old coverage, a critical piece from 2021, a competitor comparison the marketing team had never seen. Everyone who reviewed the audit was surprised during our Teams call.

She looked into her camera — at nobody and everybody at once — and said, almost matter-of-factly: “AI can’t quote coverage we never generated.”

She’s right, and most boardrooms haven’t properly sat with what that means yet.

Ask an AI assistant about your company right now. What probably comes back is a patchwork: a review site, an old article, a forum thread from 2022. Assembled with complete confidence. Presented as fact. Most people reading it will believe it.

Two peer-reviewed studies from 2025 tested this directly. Citations in AI responses increase user trust, even when those citations are completely wrong. Researchers found that responses with random, irrelevant citations were still rated more trustworthy than responses with no citations at all. Most participants never checked. Among those who did and found the citations were nonsense, trust dropped, but only then.

The default, for most people, is to believe. AI will describe your brand with complete confidence whether you’ve earned that description or not.

Generative AI systems don’t surface whoever publishes the most. They cite whoever is most corroborated — consistently — across credible third-party sources. AirOps analysed over 21,000 brand mentions across ChatGPT, Claude, and Perplexity, and 85% of the citations came from third-party sources. Brands were 6.5 times more likely to be cited through earned media than through anything they published themselves.

The University of Toronto studied citation patterns across industries — consumer electronics, automotive, software. The answer was the same everywhere. Earned media dominated. Owned content barely registered. The brands publishing the most weren’t the ones being cited.
SEO got you here. It won’t get you there.

Traditional search rewarded whoever ranked highest for the right keywords. AI rewards whoever has built the deepest, most consistent trail of independent validation. Those aren’t the same thing. The rules changed while most brands were still perfecting the old playbook.

The standard response is to publish more. Post more. Build the content engine.

The data doesn’t support it. Ahrefs looked at approximately 14 billion web pages and found that 96.55% receive zero organic traffic from Google. Zero. HubSpot, probably the most sophisticated content operation in the world, watched its blog traffic fall 75% in two years as AI Overviews began answering the questions their content was written to rank for.

If the content marketing playbook isn’t working for the people who wrote it, it isn’t working.

AI systems are trained on what already exists. Right now, they’re learning which brands are credible, which sources to trust, which names to cite. Those patterns will be hard to shift once they’re set. Seer Interactive tracked 3,119 search queries across 42 organisations over 15 months. Brands cited in AI Overviews earned 35% more organic clicks than those not cited. Brands not cited saw click-through rates fall 65% year on year.

Gartner predicts PR and earned media budgets will double by 2027 as organisations work out that AI visibility is built through independent credibility, not owned content volume. I run a company that benefits if you believe that. The data is worth looking at anyway. The smarter brands already have.
None of this is an argument for abandoning owned channels. Your blog, your social presence, your press releases — these are raw material. They are not the credibility. They are inputs into a system where what others say about you matters far more than what you say about yourself.

The brands winning in AI search aren’t the ones with the biggest content teams. They’re the ones with the deepest independent coverage, the most consistent third-party validation, and the richest citation trails built over time.

It’s what good comms people have been doing all along.

The comms leader on that call saw it immediately. She’d spent years making the case for coverage that couldn’t be tracked, relationships that didn’t show up in dashboards, credibility that compounded slowly and invisibly. She lost that argument once already.

The metrics have caught up. Whether the boardroom has is another question.

Image: Getty Images/Unsplash

The pattern we should’ve seen coming

OpenAI is building an ad platform for ChatGPT. Google’s sliding “sponsored” links into AI search results. Microsoft beat them to it with Bing Chat.
Sound familiar? It should. It’s the same playbook search ran 20 years ago: free access first, monetization later.

Every major platform eventually sells what it organizes. Search organized information and sold attention. Social media organized connections and sold influence. AI organizes trust, not by judging it, but by scaling what the web has already learned to trust. This time, the product isn’t attention — it’s credibility.

The new gatekeepers

Here’s what most people miss: Language models don’t use just any source to create responses. They digest what exists — and, crucially, what they’ve learned to trust.

Early audits show something uncomfortable: AI assistants lean heavily on established outlets — CNN, BBC, The New York Times. The top ten capture nearly 80 percent of citations. Everyone else fights for scraps.
But there’s a bigger problem — those credible platforms are vanishing.

The trust vacuum

Journalism is bleeding while AI feeds. Google’s AI summaries alone could cut publisher traffic by 25 percent — $2 billion in lost revenue.

In the U.S., ad spend on news has fallen by nearly a third since before the pandemic and is four times less than it was in 2013. More than 60 percent of marketers still avoid “hard-news” placements altogether, treating real-world reporting as brand-unsafe. Keyword blocking of terms like war, politics, and climate costs U.S. publishers another estimated $2.8 billion a year, according to the University of Baltimore.

Yet there’s a flicker of reversal. Data from DoubleVerify shows ads on news sites drive nearly 10 percent higher engagement than non-news content. Some agencies say clients are reentering news-driven channels — even Fox and CNN — because audiences still trust those environments more than the social media feeds that replaced them.

After years of retreat, brands are rediscovering a basic truth: Reliable context is the real brand-safety filter. Avoiding the news may protect optics in the short term, but it erodes the very ecosystem that makes credibility possible.

Over 10,000 journalism jobs have vanished in the past 3 years. The Los Angeles Times, NBC News, CBS News — gutted. Fewer reporters mean fewer facts checked, fewer sources verified, and fewer reasons for audiences — or algorithms — to trust what they read.

The awkward saviors

Corporate funding now reaches roughly half of U.S. newsrooms — a survival tactic that demands clearer guardrails to protect independence.

The experiments are everywhere — and each carries tension. Tech giants fund fact-checking initiatives while being fact-checked themselves. As ad dollars retreat, surviving outlets turn to the same brands whose marketing budgets once sustained them.

Done wrong, it looks like manipulation. Done right, it’s something else entirely — business investing in the infrastructure of truth.

But what does “right” actually mean? We’re defining those rules in real time, with AI systems watching and learning from every decision. Transparency is the difference between earning trust and buying silence.

When business became the adult in the room

Here’s a number that should make journalists pause: the 2025 Edelman Trust Barometer finds business remains the most trusted institution globally at 62 percent — ahead of NGOs (58 percent), government (52 percent), and media (52 percent).

When the public trusts CEOs more than journalists, it’s flattering and terrifying in equal measure. Because when business becomes the most trusted voice, communications stop being marketing. They become part of the information infrastructure society relies on.

How to lead in the age of AI trust

Forget gaming the system. Here’s what actually builds credibility with people — and with the machines now mediating them.

Radical clarity. Publish facts, not spin. Make claims others can verify. Companies that admit challenges alongside successes build authority that lasts.

Consistent verification. Share information where scrutiny still exists: peer-reviewed research, regulatory filings, and credible newsrooms with editorial standards. AI systems are already learning to weigh these sources differently from corporate blogs.

Invest in the ecosystem. Support legitimate outlets through advertising or partnerships, but do it transparently and protect editorial independence.

The tools already exist

None of this is new. Commercial newswires have been part of the trust infrastructure since the 1950s, when PR Newswire and Business Wire became how markets moved. Stock exchanges trusted them. Analysts quoted them. Their structured releases fed the databases that became search algorithms.

At 3BL, we inherited that architecture. Sustainability data now travels the same pathways as financial disclosures carved decades ago. What markets once ignored as “non-material” now drives investment, regulation, and reputation.

The lesson: Trust compounds. Decades of structured, verifiable information built the credibility that AI systems now index — and amplify. Or erase, if we drown it in noise.

The choice

Here’s what could be coming: AI platforms are already experimenting with credibility tiers — ‘verified sources,’ ‘reliable domains,’ whatever they’ll call them. Information from outlets with consistent editorial or third-party verification will rise. Everything else will sink.

That doesn’t mean companies should publish less. It means they should publish differently.

If your releases, reports, or updates can be traced through citations, filings, or credible coverage back to independent verification, AI will treat them as a signal. If they only point back to you—repeated, reworded, or syndicated across your own channels—the systems will learn to classify them as noise. You’re not punished for communicating. You’re punished for circular sourcing.

Every communications budget now does one of two things: It funds trust or floods the pool.

The smart brands are already keeping the water clear — quietly, consistently, before anyone starts charging admission.

And the irony? AI might finally force PR to rediscover its purpose: proving what’s real. Not spinning it. Not amplifying it. Proving it.

Is your content signal or noise? The age of AI trust is already here. 3BL helps communications leaders make your information traceable, citable, and credible — before the algorithms start deciding for you. [Learn how trusted distribution works →]