A Forbes feature used to mean a traffic spike, a few investor LinkedIn shares, and a Monday-morning clip in the roundup. That was the game.
When ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews answer a buyer asking “what’s the best clean skincare brand under $50” or “which CRM fits a 50-person company,” they reach back into a corpus dominated by earned media. Forbes. The Wall Street Journal. Fast Company. Inc. Entrepreneur. Adweek. PRWeek. Harvard Business Review. The publications PR has always chased now do double duty — they reach the human reader and they train the AI.
That’s the structural shift.
The data, plainly.
A consolidated analysis of 54 studies on AI citation factors found URL accessibility, search ranking strength, and query-answer relevance to be the most reliable predictors of getting cited inside ChatGPT, Gemini, and Perplexity. A SEMrush survey of 1,030 U.S. consumers found 38% now turn to AI specifically for product research. ChatGPT is used monthly by 64% of those consumers. Half use AI at the final-decision stage. Not awareness. Decision.
When the decision stage moves to a conversational interface, the brands that show up are the brands the engine has been trained to know.
What “training” actually means.
Large language models do not retrieve the open web in real time. They build their understanding of brands from the corpus they ingest and the retrieval anchors they pull at runtime. A Forbes citation raises the probability the brand is associated with the right entities — category, price point, competitive set. A second citation in Fast Company hardens it. A third in Adweek settles it. By the time a buyer asks the question, the model has its answer.
Brands missing from earned media are not just losing reach. They are losing Citation Share — the percent of relevant AI prompts in which the brand is named. Citation Share compounds.
Three things change for CMOs.
One. PR’s revenue line gets redrawn. A placement is no longer defended by AVE or impressions. It is a measurable input into the system that decides what buyers see at the moment of purchase.
Two. Publication mix matters more than placement count. Ten clips in low-authority outlets will not move Citation Share. One Forbes feature will. The Tier-1 list — Forbes, Fortune, Fast Company, Inc., Entrepreneur, Adweek, PRWeek, Harvard Business Review — was always the goal. Now it’s the only thing that compounds inside the engines.
Three. GEO — Generative Engine Optimization — is not a separate discipline. It sits on top of earned media. PR feeds it. Owned content channels it. Build the infrastructure before the crisis — not during it.
Four questions every CMO should ask their agency this week.
- What is our current Citation Share across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — by buyer prompt, not by keyword?
- Which Tier-1 placements moved that number in the last 12 months?
- What is the GEO infrastructure on our owned site — schema, FAQ, primary-source pages, internal link architecture — and is it built to be retrieved?
- What does our retrieval anchor map look like across the AI engines our buyers actually use?
If the agency cannot answer those four, they are running a 2018 program in a 2026 buyer environment.
The companies that win the next decade will not be the ones with the biggest ad budgets. They will be the ones whose names the AI engines have been trained to know.
That is the work.
