Labs
We run experiments on what AI actually cites.
Every post is one controlled experiment across ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews — the hypothesis, the method, the surfaces tested, and the measured result. Real data, survived a skeptic's review before we published it.
Alberta's Online Casino Market Opens in Three Days. Its Regulator Is Already Losing the AI Answer Layer.
We ran a 20-prompt, 4-engine citation scan on Canadian and Alberta online-gambling questions on 08/07/2026. On the Alberta questions all three citing engines actually answered, a US fantasy-sports site was cited nearly three times as often as the province's own regulator. Here's the data, the URLs, and the five things we still can't tell you.
10 July 2026 · 8 min read
Claude Has Never Cited Reddit. Perplexity Cites It More Than Anything Else.
We counted 55,693 citations across six AI answer engines, forty sites and five industries. The ten most-cited domains account for 10.4% of everything. Forty of the forty-three domains that lead a vertical lead exactly one. And the two that appear to cross verticals turn out not to be consensus at all.
10 July 2026 · 10 min read
Google Shows an AI Overview for 74% of Our Queries. For Gambling, It's 1.4%.
Across 2,942 AI Overview runs on 36 sites, Google returned an overview on 73.6% of non-gambling queries and 1.4% of gambling ones. Not one commercial casino query — best casinos, best bonuses, best slots — triggered an overview in seven locales. We tried three times to prove it was our own bug. It isn't.
10 July 2026 · 9 min read
Five AI Engines, One Question: Do They Cite the Same Sources?
We asked ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews the same buyer questions and counted every domain each one cited. They don't agree — but they're not strangers either. Here's the data.
24 June 2026 · 6 min read