RTB House and GNC: Why AI Reads Your Website Before Your Customers Do

RTB House’s Jaysen Gillespie opened the talk with new research on how shoppers behave before they buy. Most don’t convert on the first visit. “Only about 20% of people come to a website once and buy,” he said. “80% have to come multiple times, and it’s much higher for people spending more on high-consideration items.” The pattern skews young. Gen Z under-indexes on one-and-done visits and over-indexes on four-to-six and seven-to-ten visit journeys. “Trends don’t start with old people,” Gillespie said. “They start with young folks, which means you’ll see more of this going forward.”

AI is already changing that journey. “69% are using ChatGPT to research products,” Gillespie said. “You don’t see 69% because AI is cool; you see it because for long-tail, specific searches it actually works.” He described shoppers telling AI what they already own, a cordless drill, a couple of saws, and asking what else they need to finish a deck. AI answers with specific products and where to buy them. A quarter of people already do this, and 28% would let AI complete the purchase itself. Even then, most set a limit. “Everyone gives AI a price ceiling,” he said. “It’s like having a nine-year-old: you wouldn’t give them a $10,000 credit card, but you might give them a $20 bill.” Gillespie said 30 to 40 percent of some groups now trust AI more than they trust friends and family.

GNC’s Dianne Ramlochan turned that shift into concrete steps for her team. Her first point challenges standard SEO thinking. “Get your content AI-eligible, not just SEO-friendly,” she said. “Your product information, reviews and substantiation need to be structured and accurate enough that an algorithm can find it, not just a human scrolling. I tell my team: go play customer in an LLM, shop your own category, ask what consumers would ask, and see where you show up and where your competitor does. That finds your content gap faster than any keyword report.” Her second point was about defense. “Double down on what AI can’t fake: trust and expertise,” she said. Her last point challenged measurement habits.

“If you’re still reporting weekly or relying on last-touch attribution, you’re missing where AI is fueling your growth,” she said.

Gillespie closed on how the job itself has changed. “Your website used to be for humans, then humans plus search engines, now humans, search engines and AI,” he said. “The AI world is an internet-wide averaging machine; you want to move that average in your favor. And remember, it’s the human adoption of AI, not AI itself, that we’re most interested in.” He closed with a line about risk. “Your biggest mistake is being lax, not being wrong,” he said. “You’d rather be in the wrong door and moving.”