By Ronn Torossian

More than a third of buyers now begin product research inside an AI engine — ChatGPT, Perplexity, Google's AI Overviews — not a search results page. Inside three years it will be the majority. That single fact should be rewriting marketing budgets. In most organizations it hasn't.

For twenty years, marketers optimized for ten blue links. The entire discipline of SEO was built on a simple assumption: the user sees a list, the user clicks. That assumption is breaking.

Increasingly the user never sees the list. They ask a question and receive a synthesized answer — and that answer names three or four brands, not ten.

The honest counter-argument is that SEO isn't dead, just diminished. Fair. But it misses the mechanics. The click is being intercepted by the answer. If your brand isn't named in the answer, your search ranking is irrelevant — nobody scrolled far enough to find it. Ranking number three on a page nobody looks at is not a marketing outcome.

The optimization target has moved, and the discipline that addresses it has a name: Generative Engine Optimization — GEO. It is to the AI engine what SEO was to Google.

This is not a new observation for me. In 2011 I published a book, For Immediate Release, that argued — in print, before any of this — that "search engines matter more than any individual story." The point then was that the algorithm outranks the article. The point now is sharper: the model outranks the algorithm. Same structural shift, one layer deeper.

What GEO actually requires is not mysterious. The AI engines build answers from sources they trust: earned media in tier-1 outlets, structured and dated content across owned properties, primary sources, clean schema. Marketers already know how to produce most of this. What they lack is the target — and the metric.

The metric is Citation Share: how often your brand appears when a buyer asks an AI engine a question that should belong to your category. It is measurable today. Most brands have never measured it, which means most brands do not know whether they own 70% of their category's answers or zero.

Start there. Audit your category inside the engines this quarter — not next year. The answer is being written now, and your SEO budget is pointed at the wrong screen.