Signal Over Noise

The whole argument in one diagram

Executive summary: The mosaic mirror

The problem isn’t AI. The problem is that AI made the cost of an incoherent brand impossible to hide.

For 20 years, cheap paid traffic and unearned organic rank subsidized weak brand strategy. A company could grow with a fragmented brand because the channel economics were doing the work the strategy never did. That subsidy is over. AI now reads your brand across every surface a buyer touches and reconstructs a single picture.¹ If the picture is clear, it names you. If the picture is fragmented, it describes your category and moves on.

“Brand marketing is 10x more important under LLMs. 15x more important under Agents.”

AI added a second reader to marketing’s job: the engines that route human buyers before those buyers arrive. Two readers, one job. The mechanics are the same for both, because the signals a human trusts are the signals an AI extracts. This isn’t new. Sixty years of brand research, formalized by the Ehrenberg-Bass Institute as mental availability, has shown that memorable brands are the ones buyers think of when they’re in-market for your category.² AI extends the same mechanism to machine availability: the likelihood an engine surfaces you when a buyer prompts it. Design the brand for both, and one investment produces both outcomes.

Stripped of jargon, marketing in this era reduces to two jobs every company has to do well. Define the signal: decide what claim you’ll be known for, which buyers you’re built for, and what point of view you’ll champion. Make stuff about it: produce the content, listings, ads, coverage, and third-party mentions that carry that signal consistently across every surface a buyer or an engine encounters.

Skip the first job, and the second produces noise instead of signal, because no upstream decision aligned it. AI reads that noise and averages your brand into the category. Ninety to ninety-five percent of what AI uses to describe a brand comes from third parties, a pattern documented by both McKinsey and a University of Toronto study of AI search engines.³ If your own signals don’t add up to one story, third parties won’t repeat one either.

This isn’t reserved for big consumer brands. It’s now the structural prerequisite for any company that wants to be found, remembered, and recommended in AI-mediated discovery. The brands that win the next decade will be the clearest, the most coherent, and the most corroborated. AI builds the shortlist. Your signals decide whether you make it.

Contents

1. Part I · The shift
2. 1What AI sees that Google never didGoogle ranked your pages. AI reads your brand.
3. Part II · How AI reads brands
4. 2How brands live in memory, and how AI mirrors itThe signals a human trusts are the signals an AI extracts.
5. 3Why AI fact-checks your storyIf AI can’t verify your claims against what others say, it won’t recommend you.
6. 4Your website’s new jobEvery third-party reference anchors to one point of verification: your website.
7. Part III · Define the thing
8. 5Define the strategy behind the signalAI can only recognize the competitive difference your team can clearly articulate.
9. 6Three phases from noise to signalYour brand signal has to be defined in a room, by the people who live it.
10. Part IV · Make stuff about the thing
11. 7How to build and compound the signals AI picks upDefine the signal once; the flywheel deploys it across every surface.
12. ·On truthThe signals that make a brand memorable to humans are the signals that make it legible to machines.