 |
| Lauren Hill |
There’s a question worth asking before your next content investment: What does artificial intelligence say about our brand when a buyer asks?
You don’t have as much control; you don’t get to wordsmith your product page or tweak a press release, so everything lands perfectly. You’re at the mercy of the LLMs.
When a senior decision-maker types your category into ChatGPT or Google AI Overviews as a first-stop research filter—and nearly three-quarters of B2B buyers now do—the answer they get is constructed from … where?
A wide range of sources, some totally alien to you. And in roughly one in three cases, according to PAN’s own hallucination research, that answer contains fabricated or misattributed citations.
That’s the uncomfortable truth sitting underneath the current conversation about AI and B2B branding. Companies have spent years building content engines designed for human readers and for traditional search engines. AI has changed who—and what—is reading first.
The citation economy is here (and booming)
PAN’s “2026 Brand Experience Report” analyzed more than 11,000 links cited by ChatGPT in response to queries from B2B technology and healthcare executives. The source breakdown tells a story that every—yes, every—communications and public relations leader should sit with.
Nearly half (44 percent) of the links AI cited came from PR-influenced sources: earned media coverage, analyst reports, industry forums, reviews and social platforms. Another 30 percent came from owned properties, such as corporate websites.
| This article is featured in O'Dwyer's May '26 PR Firm Rankings Magazine |
This means 74 percent of the sources shaping AI-generated answers for B2B brands sit within the direct reach of PR, communications and marketing teams.
“So what?” you may be asking. “Why should I care about this study?”
Because we have a problem: AI adding an opaque layer to the B2B buying journey.
And this study may reveal the—hotly debated—solution: We can influence what AI cites. We can’t control hallucinations, but we can play a primary role in informing what AI says.
Specifically, the earned media slice—the editorial coverage in outlets like Harvard Business Review, Forbes and trade pubs B2B buyers trust—is the most instructive here. It represents 17 percent of total citations by volume. Smaller than owned and aggregated sources, yes. But it’s the category AI draws on most heavily when it needs to form an opinion about a brand. It’s the category that confers authority. And it’s the category that a PR program can directly impact.
When the infrastructure isn’t there, AI fills the gap
So, what happens in the absence of that citation infrastructure? Well, AI hallucinates.
As noted above, our research found that 31 percent of the links ChatGPT cited for senior-level B2B queries were either misattributed or completely fabricated. Notably, AI reaches for authority—credible-sounding analyst reports, recognizable publication names, senior executive quotes—and when the verifiable source isn’t there, it invents one that sounds like it should be.
What makes this especially insidious: MIT researchers found in January 2025 that AI models use more confident language when hallucinating than when stating facts. They were 34 percent more likely to use phrases like “definitely” and “without doubt” when generating incorrect information. The more wrong the AI is, the more certain it can sound.
Let me add another wrinkle. Historically, if someone posted something inaccurate about your brand, you could track it via a backlink tool and reach out with a correction. That was annoying, for sure. But it wasn’t as nebulous—or potentially damaging—as what AI can do today.
In a zero-click discovery environment—where a meaningful share of buyers skip your website entirely because the AI summary gave them what they needed—errors shape perception before your content is ever read.
Some examples of this hallucination may include a fabricated quote attributed to your CEO, a competitor credited with your research, or a report that doesn’t exist, attached to your brand name.
The downstream effect is a weakened credibility foundation at precisely the moment buyers are relying on AI to help them evaluate options quickly.
Earned media is AI optimization strategy
The reframe that our research demands is this: Earned media is no longer only about reach and awareness. It’s citation infrastructure. Every placement in a credible outlet, every analyst mention, every contributed byline in a publication AI trusts … these are inputs into the AI-generated answers buyers will receive when they research your category.
Brands that treat their PR program as a credibility-building engine are doing two things at once:
- Building trust with human readers.
- Giving AI something accurate and authoritative to draw on when it talks about their brand.
The brands that don’t build that infrastructure aren’t necessarily at risk of being overlooked. (Though that is a common consequence.) The real issue is that they will get described by AI anyway, with whatever the LLM can find or invent.
The AI in PR effect: same work, higher stakes
None of this is new work. PR professionals have always understood that credibility is accumulated. What’s new is the stakes. Buyers now use AI systems that are synthesizing your earned media footprint before a human ever sees the output. Case in point: Forrester found that 61 percent of buying journeys are considered “complete” before a buyer even contacts a vendor. That number rises when AI offers comparisons.
That means our discipline matters more. As branding pros, we need to establish …
- Consistent, verifiable presence in the outlets and channels AI trusts.
- Proof points that are documented and attributable.
- Research and frameworks that have your brand’s name on them in places AI can find.
The brands that will be represented accurately in AI-generated answers are the ones that have given AI something accurate to represent them with. Again, that’s not a new mission for communications teams. It’s the oldest one executed with new urgency.
***
Lauren Hill is VP of Marketing at PAN.
The holding company model is in structural decline. Here's what comes next.
AI visibility is becoming a top priority for corporate executives, leaving CMOs and VP-level marketers to scramble for the right way to ensure that brands stand out in AI-driven search and LLMs.
Today’s digital media landscape has reshaped influence and information discovery, which requires PR programs to adapt by changing their approach to perception and brand visibility.
The traditional press had already lost much of its status as a gatekeeper of information before AI came on the scene—and now its position is even more tenuous. In this brave new world, brands and communicators increasingly need to tell their own stories.
If the work that once built foundational skills—creating content—is automated, where does the next generation of communicators develop judgment about what effective content looks like?
