Buyers no longer always search, click, and read. Increasingly, they ask ChatGPT, Claude, Perplexity, Gemini, or Google AI Overviews. The brands named in those answers enter the consideration set. The brands omitted from those answers may never be seen.
That changes what public relations is for.
The new audience is the machine
AI engines now sit between brands and buyers. They decide which companies get cited when someone asks a category-defining question, a comparison between vendors, a "best of" query, a question about who leads a market. A PR team that doesn't show up in those answers loses reach it once took for granted.
The press release didn't die. It stopped being read by humans first. It is increasingly read by retrieval systems that decide which brands enter the answer, and which don't.
What this replaces, in practical terms, is impressions. What replaces it is citation share, a working term, used loosely across the industry, for how often a brand appears in AI engine answers to the prompts that matter in its category. The vocabulary is still settling. The shift it describes is not.
The opportunities
A few moves are already producing measurable gains:
Earn the citation. AI engines disproportionately surface tier-1 earned media, Forbes, Fortune, the Wall Street Journal, Reuters, and the trade leaders in each vertical. A single placement in one of those outlets can influence recurring citations across the major engines. A press release on the wire, on its own, does very little.
Build retrieval anchors. Brands that publish original research, structured data, and entity-rich content tend to surface in answers more reliably than brands that publish category essays. Across the AI visibility studies Everything-PR has published, the pattern repeats in every vertical reviewed, original primary data is pulled into answers at meaningfully higher rates than commentary.
Measure what you used to guess. Citation tracking across the five engines is now operational. Brands can see, by query, by engine, and by week, whether they are gaining or losing share of the answer. The PR function that ran on impressions for two decades has, for the first time, a metric that ties more directly to revenue.
The challenges
The honest constraints are not the ones most agencies highlight. Ethics, bias, and overreliance are real but well-rehearsed. The binding constraints sit elsewhere:
Speed of category formation. AI engines anchor on the first authoritative sources they crawl. Late entrants compound disadvantages. Brands that wait twelve months to build citation infrastructure end up bidding against brands that started in 2024.
Discipline collapse. Earned, paid, GEO, influencer, and content used to be five separate functions with five separate measurement systems. AI engines do not care. They cite whatever is authoritative, structured, and retrievable. Brands that organize the five disciplines as one operating system are pulling ahead.
Loss of attribution control. A buyer who gets the answer inside ChatGPT often does not click through. The brand never sees the touch. PR teams measuring on referral traffic are looking at a smaller fraction of the buyer journey than they used to. The useful metric is moving upstream, toward the citation itself.
The playbook
Audit citation share across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. Baseline first.
Identify the prompts that should belong to you, the buyer questions where your brand has a defensible claim to be in the answer.
Map the sources the engines cite for those prompts. Optimize toward those publications.
Publish original research at a regular cadence. Generic commentary tends not to move the metric.
Track weekly. Citation positions shift. So does category competition.
The brands that execute this consistently in 2026 are most likely to own the category answers in 2027. The ones that do not risk becoming invisible in the layer where buyers increasingly make decisions.
The bottom line
PR didn't get a new tool. PR got a new audience. The audience is the machine that answers before the human reads, and the discipline rebuilds around that.
The Ethical Questions Facing AI in Marketing and PR · How to Measure Citation Share Across ChatGPT, Claude, Perplexity, and Gemini
