Six experts. Six myths about how search engines and AI systems actually work, and the data that breaks each one.
Most of what teams believe was inherited, not measured. It sounds right, it gets repeated in threads and on stage, and almost nobody checks it against their own data.
So we asked six people who did.
Each episode: one expert, one myth, one dataset, under 90 seconds.
01
Prompt tracking is not full AI visibility
Kevin Indig, Growth Memo
Ask ChatGPT, Perplexity, or Gemini the same question twice, and you’ll often get two different answers. That randomness is exactly why fixed prompt sets can’t tell you how visible your brand really is across search engines and AI systems.
Kevin Indig (Growth Memo) breaks down what’s missing from prompt tracking and what to add if you want a number you can actually trust.
02
llms.txt isn’t for AI visibility, it’s for agents
Crystal Carter, Wix
Google is indexing more llms.txt files right now than sitemap.txt files, while most SEOs are still arguing about whether the thing even works.
Crystal Carter (Wix) reveals what llms.txt is actually for, and why Google, Stripe, and Claude are already treating it as core infrastructure for AI agents.
03
Answering the prompt is not the same as winning AI search
Jérôme Salomon, Oncrawl
Every AI prompt quietly branches into dozens of related questions before an answer ever gets generated, and most content strategies are only built for the one question that was asked.
Jérôme Salomon, Product Lead at Oncrawl, breaks down query fan-out and what Oncrawl’s data reveals about the content AI systems actually reward.
04
If it’s not in the HTML, AI can’t see it
Chris Green, Senior Consultant
GPTBot, ClaudeBot, and PerplexityBot all download your JavaScript files. None of them run a single line of it.
Chris Green breaks down what that means for content injected client-side or buried in images, and why “AI crawlers act like Googlebot” might be the most expensive assumption in your tech stack.
05
Getting retrieved isn’t getting cited
Pieter Serraris, OMcollective
Your logs show AI systems crawling the site constantly. Almost none of that traffic tells you whether a single page made it into an answer.
Pieter Serraris (OM Collective) breaks down the gap between getting crawled and getting cited, and why counting bot hits in your logs might be measuring the wrong thing entirely.
06
A checklist isn’t a process
Kelly-Anne Crean, Koozai
A single noindex tag survived a full technical SEO checklist, sat live in production for two weeks, and wiped out 40% of the site’s traffic.
Kelly-Anne Crean (Koozai) breaks down where the real failure happened and what needs to change so it doesn’t happen again.
The pattern
Read all six together and one thing becomes obvious: nothing here was unknowable. Someone just looked, and wrote down what they found.
The myths survive because checking takes longer than repeating them. So here’s the real question: how long has it been since you checked your own site?
