Spend three minutes on a Patagonia, Burton, or Allbirds site and the pitch is barely the point. The copy sounds like it was written by someone, not assembled by a committee protecting itself from risk. The brand admits something — a tradeoff, a limitation, an opinion that could lose it a customer — instead of asserting only what flatters it. Nothing about it reads like it was optimized to convert. It reads like it was made by people who'd argue with you if you got it wrong.
Most brand websites don't feel like that. They feel like a value proposition with photography. And for a long time, that was fine, because the website's only job was to receive a person who'd already decided to look.
That job is changing.
The visit you used to count on
A person asks Google a question. Increasingly, Google answers it directly, summarized, synthesized, complete, without sending that person anywhere. No visit. No landing page. No chance for a brand to make its case, because the case was never asked for.
This isn't speculative. An April 2025 Ahrefs study found that the presence of an AI Overview in search results led to a 34.5% lower clickthrough rate for top-ranking pages, and some searches are running 60 to 70% lower. Google's Generative UI doesn't link out to sources anymore. It reads them, builds its own interface, and keeps the person inside Google. Which means for a growing number of searches, the competition isn't another brand's website anymore. It's Google's version of yours.
Most organizations are still planning their digital presence as if the click is guaranteed. It isn't. That shift changes what a website is for — and it makes the difference between the first kind of brand and the second matter more than it ever did before.
"The click used to be the baseline. Now it has to be earned."
The mistake most organizations are making right now
The instinct is to treat this as an optimization problem: structure content so it gets pulled into the summary, get cited inside the answer, win visibility inside someone else's interface. That solves the wrong problem.
Being cited inside an AI answer is not the same as being chosen. A brand can be accurately summarized and instantly forgotten. The organizations that keep earning the visit won't earn it by being the most legible source for a machine to quote. They'll earn it by being the kind of brand someone chooses to see for themselves, even after the system has already answered the question.
That's the part a generative interface cannot touch. It can extract facts. It cannot extract the choice to sound human when sounding safe would have been easier. It cannot replicate what it feels like to land on a site that clearly believes something — the warmth of a brand that sounds like a person rather than a department. One that'll tell you what it won't do, not just what it will.
Patagonia, Burton, and Allbirds didn't build that quality for this moment. They built it over years, holding a real position long enough that you don't need to take their word for it anymore. Allbirds has been telling the same material story since before sustainability was a checkbox. Burton talks like it's run by people who ride, not people who market to people who ride. That's not positioning. Positioning is confident. This is something narrower and harder to fake — it's warm. And warmth compounds in a way that a content strategy built for citability never will.
Why this is a leadership problem, not a marketing fix
This shows up first in the budget conversation, not the brand workshop. Someone proposes spending real money on tone, voice, and the slower work of building something people would actually seek out. Someone else asks, reasonably, whether that spend is justified when the AI extraction work is cheaper, faster to prove, and easier to report on. That question wins more often than it should, because its return shows up next quarter and the other return doesn't show up at all until a competitor's site starts taking the visits that used to be automatic.
That's the actual choice, not whether AI search matters. It does, and ignoring it is its own mistake. The choice is whether an organization is willing to fund the part of the work that doesn't show up in next quarter's traffic report, on the belief that it's the only part a machine can't eventually replicate.
The click used to be the baseline. Now it has to be earned. Most brands are spending to be found and assuming that's enough. It isn't, and the ones that figure that out last will have the clearest view of what they should have built instead.
Looking for more thought leadership?