For thirty years the press list was a guess. Now we can measure it. The list is not what you think.
AI Communications, defined
AI Communications is the discipline of building authority across AI answer engines — ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — alongside earned media, digital, and influencer channels. It combines public relations, Generative Engine Optimization (GEO), and AI-visibility research to influence the answers anyone asking now begins with.
For thirty years, every public relations agency in America operated on a shared myth: that we knew which publications mattered.
The myth was useful. It gave clients a reason to buy retainers. It gave junior staff a target list to pitch. It gave senior practitioners a way to explain why a Wall Street Journal hit was worth more than a trade pickup.
The myth was also a guess.
There was no way to measure which publications actually drove buyer decisions in a category. There was no way to measure which publications were read by which audiences. There was certainly no way to measure which publications would still be cited years later when a buyer asked a question.
So we did what any industry without data does. We told each other the answer and called it consensus.
The tier-one media list is dead. AI engines made it measurable. The list is not what we thought. This is the fourth canonical piece in the AI Communications series, the deep dive on Layer 3 of the Stack.
What the old list was
The old tier-one list was a shared social construction.
The New York Times, The Wall Street Journal, The Washington Post. Then the consumer marquees — Forbes, Fortune, Vanity Fair, Vogue. Then the business trades by industry — Adweek, PRWeek, Variety, Hollywood Reporter. Then the regional papers and category trades.
The list was inherited. Most agencies' tier-one definition in 2024 was a slight modification of the same list from 2010 — same publications, same priority order.
This list was not wrong, exactly. The marquee publications did move opinions, did set narratives, did open doors. But the list was incomplete. It missed publications that mattered enormously to category-specific buyers. It overweighted publications that mattered to investors but not customers. It treated B2B trade publications as second-class even when they were the only publications a B2B buyer read.
And critically — the list was static. A 2024 tier-one list looked suspiciously like a 2014 tier-one list.
What changed
AI engines train on, and pull from, the public web. The publications they actually pull from in any given category are observable. We can run the queries a buyer would run, capture the answers, and count which sources the engines cite.
That observability changes everything.
The publications the engines pull from for a beauty query are not the same as the publications they pull from for an enterprise software query. The publications they pull from this year are not the same as the publications they pulled from two years ago.
Different category. Different list. Same methodology.
And the lists are measurably different from the inherited tier-one model. The pattern is consistent across categories.
What the new list looks like
Three patterns repeat.
The "ignored trade" pattern. In nearly every category, a trade publication that the consumer-PR world considers second-tier is doing more citation work than the marquee consumer book. A B2B SaaS category dominated by Wired and Fast Company in inherited media plans is actually pulled from heavily by The Information, Stratechery, and category-specific newsletters. A beauty category dominated by Vogue and Allure is actually pulled from Beauty Independent, specific Substacks, and trade titles consumer PR teams have never pitched.
The "old interview" pattern. Interviews and features from three to seven years ago are often the dominant source the engines pull from. A 2019 Wall Street Journal feature can outrank a 2024 Forbes feature in citation frequency, because the older piece has more backlinks, more secondary citations, more authority signals. This rewards depth over recency in a way the impressions-era media plan ignored.
What this means operationally
The media list is now a research output, not an inheritance. The first deliverable on any AI Communications retainer is a category-specific Citation Share audit that names the publications the engines actually pull from for the client's category. The agency that hands a client a tier-one list copied from a 2019 template is selling fiction.
The pitch list expands in some directions and contracts in others. Specialty trades, Substacks, and category newsletters move up. Some marquee consumer titles move down — not because they don't matter for prestige but because they don't move Citation Share for the buyer queries that matter.
The success metric on a placement changes. A placement in a publication the engines pull from is worth more than a placement in a publication they don't — even if the second publication has ten times the reported impressions.
Old interviews are an asset. Comms teams now need to identify the historical interviews and features in the engines' citation set and protect them. They are existing Citation Share assets.
Wikipedia and Wikidata become media work. They are publications too. They are pulled from heavily. Communications teams that have not owned their Wikipedia entries because "that's not PR" are conceding citations to whoever last edited the page.
What stays the same
Earned media still matters more than anything else. The substrate of AI answers is still publications. The discipline of earning coverage is still the discipline. The only thing that has changed is which coverage matters.
Tier-one prestige still has uses. A New York Times feature still helps with talent recruiting, investor signaling, board meetings, consumer trust. It is no longer the highest-leverage citation play, but it is still a high-leverage play in other dimensions. Stop confusing "tier-one for retention" with "tier-one for retrieval."
Journalist relationships still matter. A reporter who covers your category well will still write the piece that becomes a citation source in three years. Cultivating journalists is not over. The list of which journalists matters is different.
How to build the new list
Step 1: Define the query set. Same as the Citation Share audit. The questions a buyer in the category actually asks.
Step 2: Capture the citation sources across the five engines. For each answer, log which publications, websites, and authors are cited. Build a frequency-ranked list.
Step 3: Filter and reorganize. The top-cited sources by frequency become the new tier-one. Sort by category-relevance. Cross-reference with reach data. The list that emerges is the new media plan.
The first time most agencies run this audit for a client, they find that thirty to fifty percent of the inherited tier-one list does not appear in the actual citation graph. And forty to sixty percent of the new tier-one list was not on any pitch list the agency owned.
That gap is the discovery. That gap is the strategy.
The new tier-one
The new tier-one is not a list at all. It is a research output. It is per-category, per-client, refreshed quarterly.
It contains some of the old marquees, some of the trades, some of the Substacks, some of the encyclopedia and reference sites, some of the niche author pages, and almost always something the agency had never heard of.
It is the actual answer to the question that has driven public relations strategy for a century: which publications matter?
The answer used to be a guess. The answer is now a measurement.
Anyone still selling the old tier-one list is selling a relic.
