You optimized pages, built links, and fought for position one. That playbook isn’t dead, but it’s no longer as important because your buyers aren’t starting on a results page anymore. They’re starting in ChatGPT, Gemini, and Perplexity. And what those tools say about you is built almost entirely from what other people have published about you.
The research phase has left your website
Forrester’s Buyers’ Journey Survey of nearly 18,000 global business buyers found that 94% used AI in their buying process, up from 89% the year before. Twice as many buyers named generative AI or conversational search as a more meaningful source of information than any other source, far outpacing vendor websites, product experts, and sales.
The stakes of that early, invisible research phase are hard to overstate. 6sense’s Buyer Experience Report found that buyers purchase from one of the vendors on their day-one shortlist 95% of the time, and they don’t engage sellers until roughly two-thirds of the way through the journey. When they do reach out, the vendor they contact first wins the deal about 80% of the time.
Put those two findings together and the picture is uncomfortable. The shortlist is formed before you know the deal exists, and AI now helps write it. If you’re not in the answer, you’re competing for the 5% of deals where the shortlist gets reshuffled.
AI builds its answers from earned media
So what determines whether AI names you? Muck Rack’s “What Is AI Reading?” study analyzed more than 25 million links from ChatGPT, Claude, and Gemini responses across 17 industries and found that earned media accounts for 84% of all AI citations. Paid and advertorial content accounts for 0.3%. The numbers have held steady, with earned media ranging from 82% to 89% across three editions of the study going back to July 2025.
Paid content, where most marketing budgets sit, is functionally invisible to the systems buyers now consult first.
The old currency of SEO doesn’t transfer, either. Ahrefs analyzed 75,000 brands to identify which factors predict visibility in Google’s AI Overviews and found that brand web mentions showed the strongest correlation at 0.664, roughly three times stronger than backlinks at 0.218. A backlink tells a search engine where to go. A mention tells an AI system what to trust. For 20 years we built links, and now the machines reward being talked about.
This is an authority problem and PR is the authority discipline
AI engines lean on third-party sources because they’re solving the same problem your buyers have always had: they can’t take a vendor’s word for it. Independent coverage is the trust signal owned content structurally can’t produce. Forrester’s own research on the 2026 buying climate makes the human version of the same point. To win, providers must ensure their claims can be validated through trusted external voices.
For those of us in regulated industries, none of this is optional. A health care or financial services buyer isn’t just choosing a product. They’re choosing something they’ll defend to legal, compliance, and often a board. Third-party validation is cover, and now it’s also the raw material of every AI answer about you.
And unlike paid media, authority compounds. A campaign stops producing the day the budget stops. Coverage keeps working because it sits in the public record that journalists, analysts, and AI models read every time your name comes up, and every new placement deepens it.
How to build authority
Not with a press release when funding closes. Authority is a program, and it runs across more surfaces than most PR plans account for.
Make your executives sources, not subjects.
Reporters and AI models alike cite people with a track record of substantive public commentary. Contributed articles, expert reaction to industry news, and podcast appearances all build toward the same goal: when your category comes up, your leadership is already part of the record.
Give your experts a verifiable footprint.
AI systems assess credibility the way a skeptical editor would. Who wrote this, and why should anyone trust them? Every author on your site needs a real bio with credentials, role, and experience, plus a complete, active LinkedIn profile that matches it. Bylined content from a named expert with a checkable background carries weight that “by the marketing team” never will.
Target the outlets AI reads.
For B2B categories, trade press is often the primary source material, not the consolation prize. Audit where AI answers in your category currently pull from, and pitch those publications deliberately.
Show up where third parties describe you.
Directory listings, partner pages, industry association memberships, review sites, and analyst databases all contribute to the web of independent references AI draws on. The data shows which platforms matter most: Semrush’s 13-week analysis of 230,000 prompts found Reddit and LinkedIn among the top five most-cited domains across ChatGPT, Google’s AI Mode, and Perplexity, and Profound’s citation research found Wikipedia is ChatGPT’s single most-cited source, with Reddit leading on both Google AI Overviews and Perplexity. A Wikipedia presence, where your company legitimately qualifies for one, and a reputation worth mentioning in Reddit threads both feed the machines directly. You can’t control those conversations, but you can earn your way into them.
Publish your own news consistently.
A maintained newsroom with product announcements, partnerships, hires, and milestones gives journalists and AI systems a dated, structured record of your company’s trajectory. Press releases themselves rarely get cited, but they seed the earned coverage that does.
Publish original data.
Nothing earns citations, human or machine, like being the source of the statistic. Proprietary research gives journalists a reason to cover you and gives AI a reason to name you when someone asks a question your data answers.
Take positions.
Wishy-washy commentary doesn’t get quoted. A clear, defensible point of view does. If your thought leadership could run under a competitor’s logo, it isn’t building authority for anyone.
Be consistent, everywhere.
AI systems build an understanding of your company by connecting references across thousands of sources, and inconsistency breaks those connections. Use one company name, one boilerplate, and one way of describing what you do across your website, LinkedIn, directories, bios, and coverage. If half the internet calls you a “payments platform” and the other half a “fintech solutions provider,” you’ve split your own authority in two and left the machines to guess which one is you.
Sustain it, then measure it. One placement changes nothing, but 18 months of them changes what the machines say. Track your presence in AI answers alongside traditional coverage metrics, and treat both as leading pipeline indicators, because the buyer data says that’s exactly what they are.
The record decides
SEO taught marketers that visibility could be engineered. AI search is teaching a harder lesson that it has to be earned. The public record determines what AI tells your buyers, and PR is the discipline that writes the public record. The companies treating earned media as infrastructure are the ones the machines will recommend. Everyone else is optimizing a website human’s aren’t reading like they use to.
