Can't Buy an AI Citation
You can still buy a Google ad. You can't buy your way into ChatGPT's answer. That distinction, long suspected but rarely quantified, now has real survey data behind it. A new survey of 168 PR executives from D S Simon Media found that 44% now rank earned media as the most important channel for AI discoverability, 12 points ahead of paid media. Ninety-nine percent said increasing earned media investment matters for exactly this reason, and more than half are already putting over half their AI-discoverability budget behind it.
Why the gap is structural, not sentimental
It would be easy to read that 44% figure as PR professionals simply favoring the channel they already specialize in. The more useful read is structural: a paid ad and an AI-generated answer are fundamentally different products, built by fundamentally different mechanisms, and one of those mechanisms has no slot for a purchase order.
A paid ad buys placement on a page. Someone pays, the ad appears, and the transaction is visible and direct. An AI assistant does not serve a page at all. It serves a synthesized answer, built from whatever it judges to be the most credible, relevant sources already available on a topic. There is no placement to purchase, because there is no fixed slot being filled. The assistant is constructing something new each time, from material that already existed before the buyer ever asked the question.
That means the only way to influence what an AI assistant says about a company is to influence what has already been said about that company, by sources credible enough for the assistant to draw from. That is earned media, in the broadest sense: coverage, citations, and a public record built through actually being worth writing about, not through a media buy.
What "already been said" actually means
The phrase "what's already been said about you" understates how much history is packed into that idea. An AI assistant synthesizing an answer about a company or a person is effectively summarizing years of accumulated public record: press coverage, expert commentary, published research, prior interviews, the pattern of what that company or person has actually said and done over time. A single well-placed ad campaign, no matter how large the budget, cannot retroactively create that record. It can only exist going forward, and even then, only by actually contributing something an outlet or a credible third party judged worth covering on its own.
This is why the 99% figure, the share of PR executives who say increasing earned media investment matters, is arguably more significant than the 44% ranking number. It signals an industry recognizing, in real time, that a skill set built for influencing human journalists and human audiences has quietly become the same skill set that determines AI discoverability, because both depend on the same underlying asset: a credible public record that exists independent of any single campaign.
Why paid budgets built for search no longer transfer
Budget lines built around search engine optimization assumed a stable target: a results page, with a known structure, that a well-optimized page could climb. Paid search made that target even more direct, buying a guaranteed position at the top regardless of organic ranking. Both approaches depended on the existence of a page with slots to fill.
An AI assistant's answer has no equivalent slots. It has sources it chooses to draw from and sources it does not, and that choice is made based on credibility signals no media buy can manufacture directly. A company that redirects its entire search-optimized budget toward paid media aimed at influencing AI outputs is optimizing for a mechanism that structurally does not exist inside how these assistants work. The 52% of surveyed executives already shifting more than half their AI-discoverability budget toward earned media are not making a values-based choice. They are responding to how the mechanism actually functions.
Earned authority as the only asset AI can find
None of this makes paid media worthless. It remains effective for what it has always been effective for: driving immediate, direct visibility for something a company controls completely, like an ad or a sponsored placement. What it cannot do, and what no amount of budget will change, is manufacture the kind of independent, credible public record an AI assistant treats as a legitimate source to cite.
Earned authority is not optional in this shift. It is the only thing AI discovery actually has a mechanism to find. The businesses building that record now, through original research, credible commentary, and a consistent public position worth citing, are the ones an AI assistant will already be pulling from when the next buyer asks the question that used to belong to a search engine.
This is also why the 52% of PR executives already reallocating budget toward earned media should be read as a leading indicator rather than a lagging one. Public record takes time to accumulate regardless of how quickly a budget gets reallocated toward building it. A company redirecting spend toward earned authority today is not buying immediate AI discoverability. It is starting to build the credible history an assistant will eventually draw from, on a timeline measured in the same years it took competitors to accumulate the record they are already being cited for. Waiting for the AI discoverability data to become undeniable before acting on it means starting that multi-year accumulation later than the 52% already moving.