GEO Experiment: What Happens When You Build for AI Search Instead of Google?

I recently came across a post by Abby McAdams about Assembled and how they started closing real deals because buyers found them through ChatGPT and Perplexity, not Google. Those buyers came in already educated, already convinced, and moved faster through the buying process than people who found them through traditional search. That stood out to me because it suggested something important: if AI search is already driving revenue, then optimizing for it cannot be the same as optimizing for Google.

So I decided to test that idea myself. Not as a thought experiment, but with two live websites, launched on the same day, on the same platform, with opposite strategies. Both were built in two days on Hostinger, and both are fictional brands designed to reveal how search and AI systems behave when the content strategy changes.

The two sites

geo.searchshowdown.online

seo.searchshowdown.online

What I’m testing

The core question is simple: Do SEO and GEO reward the same kind of content? I don’t think they do. SEO is written for humans searching. GEO is written for machines to answer.

Over the next two to three weeks, I’m documenting what actually happens:

• What Google surfaces.

• What AI systems surface.

• What gets cited.

• What gets ignored.

• What seems to drive discovery for B2B and B2C businesses in 2026.

Why this matters

If buyers are already using AI search to research, shortlist, and decide, then the old assumption that “search = Google” is no longer enough. Businesses may need to think about two different visibility layers:

• one for traditional search traffic,

• and one for AI-generated answers and citations.