Tech SEO Connect · December 202527 min · Advanced
Michael King, iPullRank
Mike explains how AI search systems expand one prompt into many hidden sub-queries, retrieve passages for each, then choose which to use. He covers how to see those queries for ChatGPT and Gemini, his open-source tools for generating them, and what it means for content: rank for more of the sub-queries, write clear passages under clear headings, and think beyond your own site.
Why watch Dense and technical, but the best single explanation we've found of how AI answers pick their sources.
Tech SEO Connect · December 202528 min · Intermediate
Josh Blyskal, Profound
Josh shares Profound's latest research: more than 250 million AI answers and 3 billion citations across eight answer engines. Classic SEO metrics explained only a small share of which pages got cited. What mattered more was what the engines actually see (title, description, URL and a short snippet), how recent the content is, and matching the follow-up searches an AI runs for each prompt. He also covers which kinds of sources each engine favours, and what helps products show up in AI shopping answers.
Why watch A data-heavy update on how AI answers choose sources. It comes from a company that sells AI visibility tracking, so pair it with Mike King's talk on query fan-out.
brightonSEO · October 202519 min · Intermediate
Jérôme Salomon, Oncrawl
Using data from Oncrawl customers, Jérôme splits AI crawlers into three groups: those collecting training data, those building search indexes, and those fetching pages live to answer someone's prompt. He argues the live 'user' bots are the useful ones, because their visits work as a rough stand-in for impressions in AI answers and can be tracked over time, by page and by section.
Why watch The most concrete answer we've seen to 'how do we measure AI search?' using data you already own.