Why High-Authority Domains Are Vanishing from AI Search Results

Their AI Visibility Index, covered in Search Engine Land, finds that 9 in 10 brands follow the expected pattern: stronger traditional search authority tracks with stronger AI recall. But roughly 5% — 471 high-authority brands with strong domain ratings, healthy organic traffic, and deep keyword portfolios — barely surface when an LLM is asked which vendors to consider in their category. Meanwhile, another 4% (377 brands) overperform massively in AI answers despite modest SEO metrics. The gap between those two outliers is where the next layer of domain value is being built.

The consideration set just got smaller

Here's the mechanical shift that matters for anyone trading digital real estate. If an AI assistant compresses a category into five to ten default names when a buyer asks for recommendations, the effective consideration set shrinks dramatically — even when the Google results page shows dozens of viable options. That changes pipeline planning, partner strategy, and the content budgets that fund inbound inquiries. For a domain investor, it changes what counts as a "strong" name: it's no longer just a domain rating or a traffic curve. It's whether the brand attached to that domain makes it into the AI's shortlist.

What separates the winners from the invisible

The Fractl data points to one driver that explains most of the divergence: third-party content. Brands that showed up disproportionately in roundups, expert lists, and comparison reviews tended to be far more visible to the models because that coverage gets ingested into training data. Roughly 9% of the dataset aligned with how often a brand appeared in third-party content it did not control. Overperformers like Stripe in FinTech, Notion in SaaS, and Coursera in Education built assets that models repeatedly ingest and reuse, often through coverage that reads like neutral validation. Underexposed blue chips, by contrast, look like market leaders on SEO infrastructure but appear to have almost no independent editorial footprint once you strip away their own content.

In the sector examples Fractl published, the concentration is striking. Travel was the most concentrated category: Booking.com pulled 285 mentions, Airbnb 227, Expedia 215 — about 20% of the sector's total mention volume from just three names. HealthTech followed a similar winner-take-most shape, with Teladoc at 275 versus Amwell at 220.

What I'm tracking in my own portfolio

For me, this reframes due diligence on three levels.

First, when I'm pricing a brandable or aged domain, I'm now asking whether the keyword or brand attached to it has independent third-party coverage — not just its own backlink profile. A name with editorial pickup across authoritative sites carries more weight in AI recall than one with the same DR score and zero external mentions.

Second, I'm watching for the holding cost question. If you own a domain tied to a brand that's currently underexposed in LLMs, the AI gap could be a temporary arbitrage — a fixable visibility problem — or a structural one where the brand simply doesn't register in the training data. That distinction affects whether you hold, develop, or liquidate.

Third, I'm treating AI recall as a new line item in my valuation worksheets. It's still early data, but the Fractl finding that roughly 9% of brand visibility tracks to third-party content gives me a concrete proxy: the volume and quality of editorial mentions a domain's associated brand receives across non-owned properties.

The takeaway isn't that classic SEO is dead — 90% of brands still benefit from the last decade's work. It's that the next 10% of upside lives outside your own content footprint, in the mentions you can't directly buy. That's a different kind of portfolio discipline, and one I'm still learning how to price.