System1 published a chart showing which brand assets drive early recognition in short-form video.
I wanted to see the equivalent for AI. What brand behaviours predict whether AI recommends you or ignores you?
I ran a report on our ATTENT10N platform looking at 63 brands across 12 sectors. The top quartile scored 46% higher on AI visibility — and 75% of that gap comes from whether AI can find the brand at all.
Four things separate the top from the bottom.
Earned presence is the dominant predictor (+181%). Brands with genuine third-party advocacy — Reddit threads, review depth, press coverage, community discussion — are overwhelmingly more visible to AI than brands relying on owned content alone. AI reads communities. If no one is talking about you independently, AI has nothing to draw on.
AI narrative consistency matters more than you’d expect (+38%). When AI models agree on what a brand stands for, that brand scores significantly higher. When they contradict each other, the brand falls. This isn’t about content volume. It’s about whether the brand signal is clear enough that four different AI systems arrive at the same story.
Brand substance contributes, but modestly (+21%). Distinctiveness, creative depth, bold positioning. These matter because they’re what generates the earned presence in the first place. Substance is the cause. Earned presence is the effect.
Reviews support the story (+14%). Not the biggest factor, but genuinely independent from everything else. Product experience still counts.
On the negative side: the biggest drag is losing control of the narrative (-103%), followed by creating content that doesn’t match what consumers actually ask AI (-64%). Playing it safe within the category (-26%) rounds it out.
The pattern mirrors System1’s finding for human attention. The brands that win aren’t louder. They’re more coherent. And coherence travels through earned channels that brands can influence but can’t fabricate.