Original research
How AI decides who to name
Everyone has an opinion about how AI search picks a winner. We measure it. Each study runs a fixed set of real buyer prompts against five answer engines over a stated period, then traces every answer back to the sources that produced it. What comes out is the part nobody publishes: how concentrated the naming really is, and what the models were reading when they decided.
Free to quote, chart or republish with attribution and a link. Every study states its period, prompt count and engine list, so any figure can be traced back to what produced it. Journalists and researchers can request the underlying dataset.
How we measure it