A data-driven perspective on AI SEO, GEO, and the evolution of search. I analyse visibility in AI-generated answers, their impact on organic traffic, technical implementation, and emerging SEO recommendations… measuring and testing them before turning them into actionable advice.
On this website, an AI agent no longer has to click its way through everything. With WebMCP, it can search for articles, read the article currently open, or get my profile and contact details. And as a web analyst, I can measure when it actually uses this approach.
Does llms.txt actually make sense? For an ordinary website, probably not today. As a documentation hub for LLMs and agents, perhaps. But there is a major difference between the file existing, a crawler downloading it, and an LLM provider actually using it to make production decisions. Public evidence for the last point is still missing.
Why was this article written? I have my own AI and SEO knowledge base, and we had an internal discussion about whether we should invest in markdown for AI agents. Internally, we agreed that it does not make much sense right now, but I wanted to create a public reference point for it. Going forward, […]
We have more data since the last review (November 2025). Google AI Overviews (GAIO) and AI Mode have expanded, the EU has launched a formal antitrust investigation, and new studies show a significant drop in CTR for some queries. In short, this is a continuation of the article ” Google AI Overviews and AI Mode: […]
Notice: This method is no longer functional The method described in this article can no longer be used to reliably distinguish traffic from Google AI Overviews and AI Mode in GA4.