AI traffic, agent detection, and clean analytics — with data from our own network.
A practical brand-monitoring loop for AI answers: test the prompts your buyers ask, read the machine signals on your own site, and fix the pages AI actually reads.
Seeing 'ChatGPT-User/1.0' in access logs? That's a live person asking ChatGPT about you. How to read it, verify it's real, and find out which pages assistants fetch.
AI Overviews answer searches before the click. Here's how to measure what they actually cost YOUR site — impressions vs clicks, which queries, and what to do about it.
How Perplexity picks its citations, how PerplexityBot and Perplexity-User actually behave on your site, and the concrete changes that make your pages quotable.
The measurements behind the AI-traffic shift: how much of AI referral traffic gets misattributed, how it converts, and why the crawler share of your logs is invisible to GA4.
GEO is optimizing to be cited by AI answers — ChatGPT, Perplexity, AI Overviews — instead of ranked in blue links. What actually changes vs SEO, what stays, and the measurement gap nobody mentions.
Cloudflare can block AI crawlers at the edge with one toggle — even by default on new sites. When that's right, when it silently costs you AI search visibility, and why measure-then-decide beats block-everything.
Which user-agent tokens control GPTBot, ClaudeBot, PerplexityBot, and Google-Extended, what each block actually costs you, and why robots.txt is a request — not a lock.
GA4 filters known bots automatically — from a list that predates the AI-traffic era. What its filter catches, what walks straight through, and how to fire the tag for verified humans only.
Declare page context like Tealium's utag_data, fire tags on custom events with aitm.track(), and combine both with visitor type — conversion tracking that only counts verified humans.
Agentic browsers — ChatGPT's agent mode, Perplexity Comet, Claude in Chrome — run real browsers, execute JavaScript, and fill your analytics as 'users'. What they are, why they matter, and how to detect them.
One liquid snippet shows you the AI assistants and agentic browsers shopping your store. Here's the five-minute install — plus an honest note on which crawlers Shopify's architecture hides from everyone.
A user-agent string is a costume anyone can wear. Scrapers impersonate GPTBot and ClaudeBot to borrow their reputation. Here's how IP verification separates real AI crawlers from cosplay.
llms.txt is a proposed standard: a markdown map of your site for AI systems, served at /llms.txt. Here's what it does, what it honestly doesn't, and how to add one in ten minutes.
ChatGPT, Perplexity, and Google AI Overviews don't rank ten blue links — they pick a handful of sources to quote. Here's what actually makes a page citable, and how to check if AI engines can even see yours.
Google's Gemini fetches pages wearing two disguises: a bare 'Google' user agent, and a fake Chrome UA missing the token every real browser sends. Here's how to unmask both.
Blocking GPTBot doesn't hide you from ChatGPT users — it only opts you out of model training. Here's what each OpenAI crawler actually does, and how to decide with data instead of fear.
GPTBot, ClaudeBot, and PerplexityBot read WordPress sites every day — and never show up in your analytics, because crawlers don't run JavaScript. Here's how to see them with a plugin.
GA4 has no AI channel out of the box. ChatGPT visits land in Referral, Direct, or Unassigned depending on how the link was clicked. Here's exactly where to look, and how to fix it.
35–70% of visits from ChatGPT, Perplexity, and Gemini arrive with the referrer stripped. GA4 files them under Direct. Here's why it happens and how to measure what's really going on.