More than a third of people now begin their research — for products, for politicians, for doctors, for ideas — with an AI engine, not Google. The answer ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews return is the new front page of the internet. AI Communications is the discipline of influencing what that answer says — for any institution, person, or idea that can be named inside it.

I've said for years that PR is a mix of journalism, psychology, and lawyering — an ever-changing and always interesting landscape. In For Immediate Release, I predicted that audiences would fragment, gatekeepers would weaken, and influence would belong to whoever could reach the moment of decision directly. The medium has changed three times since. The mechanic has not.

"AI Communications is a mix of journalism, psychology, and engineering — and the audience is now the machine."

It is still journalism. Still storytelling. The story now has to be readable by both humans and machines.

The discipline rewards intellectual curiosity — the willingness to push the technology, not just follow it. To understand that a schema is part of communications now. To learn how an engine retrieves, what it weights, what it ignores. The communications professionals who own the next decade will be the ones who treat the AI engine as a counterpart to be understood — not a black box to be feared.

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."

That definition is deliberate. Read it twice. Two things matter inside it.

• It is a communications discipline — not a marketing one. The work is earned. Authority comes from being cited by sources the AI engines trust: established publications, peer-reviewed research, credentialed practitioners, primary sources. It cannot be bought into. It can only be earned and then engineered.
• It is a stack — not a tactic. AI Communications absorbs public relations, GEO, AI-visibility research, structured data, and earned media into a single operating system. Anyone selling one piece of it as the whole thing is selling SEO with a new label.

Who it is for

Not just brands. Not just buyers.

A founder being researched by a job candidate. A university being researched by a parent. A candidate being researched by a voter. A nonprofit being researched by a donor. A doctor being researched by a patient. A movement being researched by a journalist. An author being researched by a reader. A company being researched by an investor.

Any institution, person, or idea that can be named inside an AI answer has a stake in AI Communications. The discipline applies wherever attention now begins — which is everywhere attention now begins.

Why the category exists now

Three numbers explain it.

• More than a third of people begin research with AI rather than search engines, depending on the category. Consumer tech and travel run higher. Financial services and healthcare are catching up fast. That share is climbing every quarter.
• Billions of AI answers are generated daily across the major engines. Each one is an opportunity for an entity to be named — or not.
• Zero is the number of times most institutions and people measure whether they are.

That gap — between where attention now begins and what the communications industry measures — is the gap AI Communications closes. PR measures impressions and earned media value. SEO measures rankings and traffic. Neither captures the question being asked inside a chatbox at 9:47 p.m. on a Tuesday and the answer that follows.

That answer is the new front page. AI Communications is how anyone earns their place inside it.

What AI Communications is made of

Six layers. They work together, not separately.

1. Earned media — measurable for the first time

For thirty years, the press list was a guess. A 2015 media planner said tier-one was The Wall Street Journal and The New York Times and you believed it because there was no way to check. Now there is. AI engines are trained on, and retrieve from, publications — and the retrievals are observable. Query by query, we can measure which publications the engines actually pull from in a category. The list is not what you think. The trade you ignored is doing more citation work than the consumer book on your wall. Earned media isn't dying. It's becoming measurable. Which means it's becoming strategic.

2. Generative Engine Optimization (GEO)

The discipline of structuring content — language, schema, citations, entities — so AI engines retrieve and quote it. GEO is to AI Communications what SEO was to digital marketing fifteen years ago: a technical practice that becomes a strategic one once leadership understands the metric.

3. AI-visibility research

The audit layer. Which engines name the entity. Which name competitors. Which name nothing. Which sources the engines are pulling from. Where the gaps are in the authority graph.

4. Structured authority

Wikipedia, Wikidata, trade reference, glossaries, schema markup, primary research, founder bylines in publications the AI engines trust. The pieces that explain to the AI engines what the entity is and why it matters. Schema literacy is communications work now — not a separate technical specialty.

5. Influencer and creator integration

Increasingly cited. Increasingly retrieved. Especially for consumer categories — beauty, fashion, food, travel, wellness — where AI engines pull from creator content inside both training data and live retrieval.

6. Crisis response inside AI answers

The newest layer. When an AI answer goes negative — outdated facts, surfaced controversy, competitor framing — the response cannot be a press release alone. It has to reach into the same retrieval graph that produced the answer.

How it is different from PR, SEO, and GEO

A short disambiguation.

• Public relations influences journalists and audiences directly. AI Communications uses earned media as one input among several into a different goal — the AI answer.
• Search engine optimization optimizes for Google's index. AI engines do not work like Google's index. They synthesize. They cite. They omit. SEO tactics — keyword density, link velocity, page speed — are necessary but no longer sufficient.
• Generative Engine Optimization is one component of AI Communications, not a synonym for it. GEO is the technical practice. AI Communications is the strategic discipline that includes GEO, earned media, AI-visibility research, structured authority, and crisis response as one stack.

A practitioner can do GEO without doing AI Communications. No one can do AI Communications without doing GEO.

The metric — Citation Share

Every discipline gets defined by what it measures.

Public relations measures impressions, earned media value, share of voice. Search marketing measures rankings, click-through, organic traffic.

AI Communications measures Citation Share — the percentage of relevant AI-engine answers in a category in which the entity is named, quoted, or linked.

Citation Share is measurable. It is comparable across competitors. It is trackable over time. And it is the single most important number a communications team will track in the next five years, because it is the number that predicts whether anyone asking will encounter the entity at the moment of decision.

A founder with 4% Citation Share inside ChatGPT for their category will lose attention to a founder with 31% — even with more press, more traffic, and a bigger marketing budget. The AI engine is the gate. The citation is the ticket through.

What an AI Communications practice does

Five activities, in roughly this order.

• Audit. Run the queries anyone asking would run. Capture the answers across the five engines. Score Citation Share. Map the retrieval sources. Identify gaps.
• Anchor. Build the retrieval anchors — pillar pages, definitional content, structured authority, primary research — that the engines will pull from. The work has to live where the engines look.
• Earn. Run the earned media program — but pitch into the publications the engines actually retrieve, not the publications a 2015 media list says are tier-one. The list has changed.
• Test. Re-run the queries weekly. Watch the citations move. Adjust the anchor content where they don't.
• Defend. When the answer goes wrong — outdated, hostile, missing — the response is research, sourcing, and earned media targeted at the retrieval graph. Not a takedown letter. Not a press release into the void.

This is operational work. It is measurable. It is ongoing. It is not a campaign.

Where AI Communications is going

Four things to watch over the next twenty-four months.

• Multimodal retrieval. AI engines are increasingly pulling from video, audio, and image content. The Citation Share calculation will need to expand. Anyone ignoring podcast appearances and YouTube transcripts is leaving citations on the table.
• Agentic commerce. AI agents are going to make purchases on behalf of users. The agent's decision will be driven by the same retrieval graph that produces AI answers today. The entity that is not cited is not considered.
• Regulatory disclosure. Expect rules — first in the EU, then the U.S. — requiring AI engines to disclose retrieval sources and to give entities a path to correct factual errors. This will create the first formal AI-rights regime for institutions and people.
• Crisis acceleration. A bad AI answer is now a crisis vector. Misinformation, surfaced controversy, competitor framing — all of it will be encoded inside the answer engines unless and until someone reaches into the retrieval graph. Build the infrastructure before the crisis — not during it.

And this list will change. The engines will change. The retrieval graph will change. The metric will refine. Anything called AI Communications in 2026 will look different in 2027 and different again in 2028. Anyone who claims they have nailed down AI Communications has not been paying attention. The work is to keep paying attention.

The takeaway

The answer engines are the new front page of the internet. AI Communications is how anyone — a brand, a founder, a candidate, a cause, an idea — earns their place inside the answer.

PR did not die. It moved. The institutions and people who understand where it moved will own the decade.

Ronn Torossian is the publisher of Everything-PR and the author of two best-selling marketing books, including For Immediate Release.