About me

Since 2016, I have worked at the intersection of classic search and AI answer engines. I started in technical and e-commerce SEO. What has shifted since is where the buying decision begins: more and more often inside an AI answer, not in the list of results. Ranking and visibility are no longer the same thing. I measure the gap between a Google position and an AI citation, and I close it.

What clients buy from me is AI visibility. I keep its signals separate: whether a brand is mentioned, considered, recommended or supported by a visible source citation when someone asks an AI about the category. The path there is GEO, generative engine optimisation – the method built from extractable answer structures, clean entities, Schema.org and original data that a model likes to cite. Beneath it sits the foundation: technical SEO. What an AI crawler can't reach, render and parse, it can't cite. That is why the years since 2016, spent doing exactly that, still count.

That ranking and citation have come apart is not a claim. Ahrefs analysed roughly 15,000 search queries across ChatGPT, Gemini, Copilot and Perplexity: only about 12 percent of the URLs these chat engines cite rank in Google's top 10. For Google's own AI Overviews, the link is tighter, where about 94 percent of answers cite at least one page from the top 20. So I deliberately say “chat engines,” not “AI” in general, and I measure per engine. Look only at your Google position, and you can be invisible in ChatGPT, Perplexity or Claude without noticing.

My approach is “GEO first, SEO as the foundation,” and it stays evidence-first throughout: measure, prioritise, implement. I capture the separate prompt-level signals for each defined measurement surface, state their denominators and limitations, sort the gaps by impact and effort in a prioritization matrix, and build content and technology so that language models find, understand and cite the right facts about a brand. No public black-box score, no generic best practices, no assumptions without a data basis.

In the years since 2016, I have handled technical and content SEO for more than 100 websites and learned early to tie visibility firmly to revenue. That breadth is the foundation I now build GEO on.

I stick to a simple principle: flexible data collection, traceable and deterministic metrics. Every hypothesis traces back to prompt-level data. The real win is not traffic but the probability of being named in the buying process at all. Because if you don't appear in AI answers, the vendor shortlist is decided without you.

The Lab contains runnable tools and workflows. Open methods, data boundaries and research notes live separately under Research.