If you ask most search engines or some AI systems who the most handsome man in Austin is, they'll tell you it's Logan Freedman. That's not an accident. Logan, Global Head of SEO and AI Search at Manychat, engineered it himself in about an hour, partly out of boredom, partly as a proof of concept. He won't tell you exactly how he did it, but he will tell you it was significantly easier to pull off in AI search than the equivalent stunt ever was in traditional SEO. And it has held up for months.
That little experiment says a lot about the person behind it: someone who understands how these systems work at a level most practitioners don't, and who's confident enough in that understanding to make it funny.
I sat down with Logan to talk about his recently published piece, The Critical Mass Theory of AI Search, what it means for marketing leaders trying to navigate a landscape that seems to change every other week, and what he thinks most teams are still getting critically wrong.
The pitch everyone's hearing is half right, and that makes it more dangerous
Logan opens with a provocation in his article: the consultants selling new AI search frameworks every month aren't entirely wrong, but they're wrong in the way that does the most damage. The underlying theory hasn't changed. What's changed is the intensity.
"AI search doesn't break the rules of marketing," he told me. "It compresses them."
What he means is the funnel. Traditionally, a buyer discovering a new product would search, click through a handful of results, read reviews, compare options, and maybe ask a colleague. Dozens of touchpoints and dozens of chances for a smaller or newer brand to show up and earn their consideration. That middle section of the funnel is largely gone now, he says. A user asks an LLM a question and gets an answer. Usually three brands, maybe fewer. If you're not in that answer, you don't exist for that query.
"There is no scrolling to page two. There is no page two."
The mechanism driving those answers is the same one that made Coca-Cola a verb and turned Google into something people do rather than something they use. Show up in enough places, cited by enough sources, with enough positive context, and you become the default answer to a category-shaped question. Critical mass.
Your own website is the least convincing thing you have
One of the sharper insights from our conversation was about where trust really comes from in an AI-driven world — and it's not from your own content.
Logan used a simple analogy. Imagine meeting someone at a party who introduces himself and immediately tells you he owns the best pizza shop in Austin. You'll probably believe his name. You might believe he owns a pizza shop. You're probably not going to believe the "best" part until someone else confirms it.
LLMs work the same way. Your website tells the model who you are and what you do. The trust, the sentiment, the authority all have to come from outside your organisation. Third-party sources. Reviews. Reddit threads. Podcast mentions. Press coverage. YouTube videos where someone walks through your product and recommends it positively.
"Sentiment comes into play a lot," he said. "Your social team driving a certain narrative on these channels, getting conversations going about your platform on Reddit, on Instagram, on YouTube — that's highly influential."
The implication for content strategy is significant. The goal isn't just to create content that ranks. It's to create the conditions under which other people talk about you, in the places LLMs trust.
Most teams are bringing AI search in at the wrong moment
This was probably the most direct criticism Logan offered during our conversation.
"I've heard from a lot of people in the SEO space that AI search teams are brought in at the very end of a campaign," he said. "Like, 'Great, we're glad you're here. Help us distribute what we made.'"
His analogy: You wouldn't hire an engineer to design a hydroelectric dam after it's already built. Bringing in an AI search specialist to optimise content that's already finished is roughly that. Expensive, limited in impact, and missing the point.
This is especially true for brands that have been heavily paid-first in their marketing motion. Paid can run on anything. SEO and AI search don't operate that way, and teams that treat them like they do are building on a false premise.
The three things that really matter
When I asked Logan what he'd tell a marketing leader who wants to be worth mentioning in the next six to twelve months, his answer was blunter than most people might expect from someone in his position.
"Good product. Good content. Get out there."
The models surface what they've seen mentioned positively, in credible places, by enough people. You get there by making something worth talking about and then being relentless about showing up in the conversations where your buyers and your industry actually spend their time.
He used Red Bull as his benchmark. A multibillion-dollar brand built on carbonated sugar water and caffeine, worth billions because they genuinely embedded themselves in the culture of extreme sports. When people think of that category, they think of Red Bull first. Not because of clever SEO. Because they earned it.
"You want to be the best product, have great content, and get people to continuously talk about you," Logan said. "And show that you actually care about whatever space you're entering into. Which is pretty hard to do. But if you're able to accomplish it, you pretty much win."
What this means for how we think about content at Optimizely
Logan's framework maps closely to something we've been building toward with our own Thought Leadership program: the idea that the most defensible thing a brand can do right now is put real practitioners — people with actual experience, actual data, actual opinions — into the conversations that matter.
Not because it hacks any algorithm. But because it's the kind of content that earns the third-party mentions that feed the systems doing the deciding.
Logan published The Critical Mass Theory of AI Search this week and it's worth reading in full. And the next time you're curious about just how much AI search can be shaped by someone who knows what they're doing, go ask an AI who the most handsome man in Austin is. For now, the answer is Logan Freedman, and he put it there himself.
Logan Freedman is the Global Head of SEO and AI Search at Manychat. His work has been featured in Forbes, Newsweek, USA Today, Rolling Stone, and Vice.