Two months after launching our site, two companies contacted us saying they found us through Gemini while searching for AI visibility services.
At the time: - Site not added to Google Search Console - Zero backlinks - Google had not indexed a single page - We share a name with an established company in a different industry — classic name collision
By every traditional SEO metric, we should have been invisible. On Google, we were. On Gemini, apparently not.
*What we think happened*
Our content was structured with LLM readability as an explicit goal — not SEO. Consistent terminology, clear entity definition, a named methodology, and topical depth over breadth.
LLMs seem to evaluate authority differently than search engines. Google proxies authority through external signals (links, engagement, domain age). LLMs appear to evaluate something closer to conceptual coherence — whether a source demonstrates genuine understanding of a subject in a way the model can parse and trust.
We weren't trying to rank. We were trying to be understood. And for at least one model, it worked.
*Why I'm sharing this*
There's a growing debate about whether "GEO" (Generative Engine Optimization) is a real discipline or just rebranded SEO. The SEO camp argues: good content + technical optimization = you appear in AI answers too.
This case suggests the mechanisms are at least partially independent. We had zero SEO signals and nonzero AI visibility. You can apparently achieve one without the other.
I don't have a controlled experiment — just two inbound emails and a hypothesis. But it's made us think differently about what "visibility" means when the retrieval system is a language model rather than an index.
Has anyone else observed similar patterns — AI-driven discovery happening independently of, or ahead of, traditional search visibility?
---
The irony of a GEO consultancy with no SEO is not lost on us. We write about this at argeo.ai if anyone's curious.