
Pete Meyers
Principal Innovation Architect, Moz
Moz
At a Glance
- Role
- Principal Innovation Architect at Moz, known as Dr. Pete
- At Moz since
- November 2012
- Doctorate
- Ph.D. in human experimental psychology
- Known for
- MozCast, SERP feature research, and The Infinite Tail keynote
From the Lab to MozCast
On paper, he has said, the path looks odd. He coded as a kid and finished a computer science and psychology degree in 1992, when the internet was barely a workplace. Before SEO, he was a cognitive psychologist, a computer scientist, and a vice president at a web-based event startup, and then he became an SEO consultant. He joined Moz in November 2012 as a Marketing Scientist.
At Moz he built and curated the Google Algorithm History project and MozCast, which tracks the “weather” of Google's results. That daily measurement habit is the ground under the talks he now gives on AI search. Our GEO and AEO industry statistics sit beside that kind of SERP tracking.
The 85 and 15 Forecast at BrightonSEO
Hours before his closing keynote at BrightonSEO 2026, he made a forecast the rest of the industry would find inconvenient. Roughly 85 percent of search intent, he said, will land on a feature-rich results page. About 15 percent will go to AI mode.
“LLMs are really bad at navigational content. They just want to go to your website,” he said. “Maybe you're gonna go straight to AI mode 15% of the time, just making up a number, but the other 85% you're gonna go to something that's more feature-rich, more organic.” His case was not that SEO is unchanged, but that teams abandoning classic work for GEO tools are shopping for the wrong thing.
The Same Engine Under AI Mode
The keynote itself, The Infinite Tail: Keyword Research for AI, treated the line between pre-LLM organic search and post-LLM AI search as mostly artificial. In 2013, the same month Google shipped Hummingbird, the Word2vec paper made vector embeddings practical, and those embeddings, he argued, have been quietly powering organic search for 13 years.
He pointed to US antitrust testimony: to ground Gemini, Google uses FastSearch, based on RankEmbed signals, to generate abbreviated, ranked web results a model can use. The same engine, he said, feeds organic results, AI Overviews, AI Mode, and Web Guide. “This evolution has been going on the back end the entire time. Our evolution has been going on the entire time.”
Losing Clicks Without Blocking the Web
In August 2026 he rejected the idea that publishers should wall off their pages from AI crawlers. “Right now we are all losing clicks and it's frustrating, but blocking content won't turn back time,” he said. Intent, he added, now matters more than exact words, in AI engines and in search in general.
For a researcher who has watched Google's results day by day for more than a decade, the AI era is a change of surface, not of job: keep measuring what the page really shows, and keep doing the brand and retrieval work that those surfaces still draw on. Our AI crawler blocking statistics cover the other side of that argument.