System Design · Part 4 of 5

There's a quiet fear under all the AI-search anxiety.

It usually comes out as a question, and it goes something like this. If people start buying through ChatGPT, and the AI lines us up next to ten other brands, what's the point of us? What stops us becoming one more interchangeable option in very long comparison list?

It's a fair concern but I believe it's directed at the wrong thing.

The AI era rewards brand quite a lot more than search ever did, and all that slow work you've put into being a memorable brand worth trusting is about to become your biggest advantage. I'd go further than that because it may well be the only advantage left that a machine can't copy.

How the AI decides who to recommend

When an AI recommends a product it's making a bet. It's putting its own credibility on the line by vouching for you to someone who came to it for help, so it leans toward brands it can trust, the same way a person would. Authority: are you a recognised name in your category? Reputation: what your customers and the press and the creators say about you when you're not in the room. Relevance: a clear identity and a focus, or a muddle. Popularity: are people talking about you, linking to you, recommending you?

Every one of those is brand. The deeper kind, underneath the logo and the colours - the reason someone picks you and then tells a friend.

A brand started as a mark on a barrel

When goods first travelled further than the next village you had no way of knowing whether the soap from a faraway factory was any good, until it carried a mark you recognised. The mark meant "same quality as last time".

A brand is a promise.

A shortcut to trust.

That's exactly what an AI is hunting for when it sorts the trustworthy from the risky. We invented the shortcut a few hundred years ago and the machine is using the same mechanism.

Why this is good news for a premium founder

If you've built a real brand, one with genuine customer love and some earned press and a community that talks about you, then you've been quietly stockpiling the exact thing AI now prizes. Brands that do the work show up more in the data-sets these models learn from, and they get recommended more often as a result. And for a premium brand this definitely cuts your way, because your whole proposition runs on perception, trust, taste and the feeling of being worth more. That's brand. It's the thing you're already best at.

The discount-driven, race-to-the-bottom sellers have one thing to show an AI, and it's comparing one low price to another and a low price is possibly the easiest thing in the world to undercut.

A trusted brand isn't.

The stores in my own orbit that came through the last few years all had the same thing about them. Something real. A genuine product or service you can't get from the AI itself, the ability to do something for the customer, assets no one else owns and a brand people trust. Rand Fishkin found the same pattern in a study Cyrus Shepard led across four hundred sites, and his summary is blunter than mine: No amount of tactical excellence saves you if the underlying thing is something an AI can replace.

Tactics decay. The real stuff compounds. For you, it's the brand.

The AI can summarise your product and line it up against nine others, but it can't be you.

But there's a catch

Only some of what the AI believes about you comes from your own site. The rest comes from everywhere else: Reddit threads. Reviews on platforms you don't own. Press. The creator who mentioned you in passing. The forum where someone asked "is this brand actually worth it" and got three replies, one of them helpful.

So who owns those rooms?

No one you know. But you've got more sway in them than you think and it comes from the same slow work you're already doing. What people believe about you now gets built and grows in rooms full of strangers and Rand Fishkin's point is that it runs on feeling far more than on facts. The conversation about you is happening in those rooms whether your product is wonderful or not. Being good is necessary. Being talked about, accurately in the places the AI reads, is what turns good into recommended.

What I'd do now

I'd run the brand test again, and this time watch the sources. Ask the AI what the internet says about you and note which pages and threads it quotes. Where it's quoting you and getting it wrong, fix your own pages. Where it's quoting an outside source, make sure there's a clear accurate version of that same thing on a page you own, somewhere a crawler can reach.

Then keep doing the slow brand things. Reviews, press, community, being consistent about who you are everywhere you turn up. Only now, count them as AI work too.

I'm no guru on any of this and I can be wrong. But brand compounds, and that's the whole game.

Now ask yourself, what's one place people talk about your brand that you've never gone and read?

Share this article

Share to Facebook
Share to X
Share to LinkedIn

Written by

Join the conversation