Nine out of ten brands — 90% — receive zero mentions in AI-powered search results, according to a new study from Search Engine Journal that also lays out four key SEO insights for marketers trying to close that gap.

The number is blunt. Whatever traffic and visibility brands have built through conventional SEO, most of it does not carry over when users ask ChatGPT-style search tools instead of typing queries into a traditional search box. If a brand is not cited by the AI, it effectively does not exist for that user.

The finding lands at a moment when publishers and marketers are still arguing about how much traffic AI search actually sends. Brand mentions inside AI answers function like citations: they arrive with an implicit endorsement, they are visible to a single user at a single moment, and — unlike a blue link — the user rarely scrolls past them to a competitor.

What does the 90% figure actually mean?

The study's core measurement is mention share: whether an AI search engine names a brand when users pose category-type questions such as "best CRM for small business" or "top project management tools."

By that yardstick, the market collapses into two camps:

• The invisible 90% — brands the AI never names, regardless of how strong their traditional rankings look.
• The visible 10% — a narrow set of brands that AI engines cite repeatedly, capturing the recommendation slot before any click happens.

The practical consequence: the winners are not competing for position one on a search results page. They are competing to be the default answer, and there is room for far fewer brands in a spoken or written AI answer than on a ten-blue-links page.

What are the four SEO insights?

Search Engine Journal distilled the study into four key SEO insights aimed at brands trying to break into the cited 10%. The study positions these as the levers that separate mentioned brands from ignored ones in AI-generated answers.

The context matters more than the checklist. AI search engines draw on the open web — reviews, comparisons, listicles, forum threads, editorial coverage — to decide which brands deserve a mention. A brand that appears only on its own site gives the AI little to work with. A brand that appears across many independent sources gives it plenty.

That reframes familiar SEO work. Structured data, entity clarity, consistent naming, and third-party mentions are no longer hygiene tasks at the bottom of a backlog. They are the raw material AI engines use to decide who gets named.

Why should operators care now?

The 90% figure is a baseline, not a verdict. Early AI search behavior is still forming, which means mention share is still contestable in a way that classic search rankings, after two decades of consolidation, largely are not.

For marketing teams, the study's implication is direct: audit whether AI engines mention your brand at all, before optimizing anything. A brand that assumes it is visible, and is not, is spending budget against a position it does not hold.

The full set of four insights, along with the study's methodology and detailed findings, is laid out in Search Engine Journal's coverage — worth reading in full before the next planning cycle, because the gap between the 10% and the 90% is only getting easier to measure and harder to explain away.

What is the takeaway?

One number defines the story: 90% of brands have zero AI search mentions. The four insights in the study are the suggested path from the invisible majority into the cited minority. Whether that path works at scale is a 2025 question — but standing still now means competing later against brands that started building AI visibility while it was still cheap.