- Jan 20, 2013
- 3,354
- 550
Okay so here's something that's been bugging me.
Every brand I talk to right now is trying to "win" AI search the same way they tried to win Google — just publish more stuff. More reports, more indexes, more of that AI-generated "thought leadership" that reads like nobody actually wrote it. The thinking is basically: throw enough content at ChatGPT and Claude and eventually you'll show up.
I don't think that's how this works.
Muck Rack ran the numbers and found something wild — about 94% of the links AI tools actually cite are earned, not brand-owned. Not your blog. Not your press release. Earned coverage. And Similarweb found that when ChatGPT does send someone your way, that traffic converts at more than double the rate of regular Google traffic.
So think about what that actually means.
Google used to hand people a list and say "you decide." These AI tools don't do that anymore. They just tell you the answer. "Here's the brand you should look at." And that answer isn't coming from your website copy — it's coming from everything that's been said ABOUT you. Press coverage. What analysts say. Customer stories. Random Reddit threads, honestly. Podcasts you did two years ago that you forgot about.
This is why I keep saying PR isn't just a "brand awareness" thing anymore. It used to be about getting your name out there, then it became about reputation management, and now — weirdly — it's becoming an SEO play in disguise. The stuff PR has always done is exactly what these models are learning to trust.
And yet most companies are reacting to this the exact wrong way. Same playbook as always: crank out more content. More self-published research nobody asked for. More filler. More noise that, let's be honest, only shows up because the brand itself is the one citing it.
Here's the thing though — a report with sloppy methodology doesn't make you an authority. Neither does an index that's really just a headline dressed up as data. You can publish 500 AI-written articles and still have zero credibility in the eyes of these systems.
I think the brands that actually pull ahead here won't be the loudest ones. They'll be the ones other people are already talking about — press, analysts, customers, whoever. That's the stuff AI models seem to trust. Not your own megaphone.
Different game than the one most teams are still playing.
Every brand I talk to right now is trying to "win" AI search the same way they tried to win Google — just publish more stuff. More reports, more indexes, more of that AI-generated "thought leadership" that reads like nobody actually wrote it. The thinking is basically: throw enough content at ChatGPT and Claude and eventually you'll show up.
I don't think that's how this works.
Muck Rack ran the numbers and found something wild — about 94% of the links AI tools actually cite are earned, not brand-owned. Not your blog. Not your press release. Earned coverage. And Similarweb found that when ChatGPT does send someone your way, that traffic converts at more than double the rate of regular Google traffic.
So think about what that actually means.
Google used to hand people a list and say "you decide." These AI tools don't do that anymore. They just tell you the answer. "Here's the brand you should look at." And that answer isn't coming from your website copy — it's coming from everything that's been said ABOUT you. Press coverage. What analysts say. Customer stories. Random Reddit threads, honestly. Podcasts you did two years ago that you forgot about.
This is why I keep saying PR isn't just a "brand awareness" thing anymore. It used to be about getting your name out there, then it became about reputation management, and now — weirdly — it's becoming an SEO play in disguise. The stuff PR has always done is exactly what these models are learning to trust.
And yet most companies are reacting to this the exact wrong way. Same playbook as always: crank out more content. More self-published research nobody asked for. More filler. More noise that, let's be honest, only shows up because the brand itself is the one citing it.
Here's the thing though — a report with sloppy methodology doesn't make you an authority. Neither does an index that's really just a headline dressed up as data. You can publish 500 AI-written articles and still have zero credibility in the eyes of these systems.
I think the brands that actually pull ahead here won't be the loudest ones. They'll be the ones other people are already talking about — press, analysts, customers, whoever. That's the stuff AI models seem to trust. Not your own megaphone.
Different game than the one most teams are still playing.
