24 results in Data & Studies
We tested OpenAI's new flagship. It searches the web 40% less than the model it replaced, and the searches it keeps go straight to brand pages, not reviews. Your pricing page is the new front line.
We tracked 4,670 brands over four months to find which signals moved with AI visibility growth. Citation share showed a 0.41 correlation with visibility change.
We analyzed 10.7 million AI answer changes to understand brand displacement in AI search. In 90% of brand entries, at least one other brand left.
We ran 161,286 prompts through ChatGPT, Gemini, Perplexity, and Google AI Overviews and measured how much their citation pools overlap.
We ran 631,999 prompt-model pairs across 7 AI platforms, from March through June 2026, each prompt repeated a minimum of 10 times.
GPT-5.6 Sol reverses the biggest finding of the GPT-5.5 era: the site: operator is back, on up to 71% of searches. Brand citations, freshness, and answer length all moved with it.
We measured citation influence across 2,000+ brands. YouTube (3.19x), Reddit (3.17x), and Gartner (3.15x) deliver 3x lift. Your own website? Just 1.3x.
We measured ChatGPT ad penetration across 363,129 responses over 31 days. For 30 of those days, ads were a rounding error. On day 31, the surface flipped.
