Every release samples a vertical's real buyer questions across the major AI engines, resolves which brands appear in the answers, and publishes the share, engine by engine, over time.
Hundreds of prompts per vertical, stratified by intent, funnel stage, and persona, mined from how buyers actually ask, then run in replicated waves.
ChatGPT, Google AI, Perplexity, Claude, and AI Overviews, sampled independently, because fewer than 10% of AI citations overlap with classic search rankings.
AI answers churn, so the Index reports confidence intervals, not fake precision. Replicated sampling, a published exclusion log, and a raw-data sample with every release.
Up to 90% of cited sources in AI answers change over time. The Index is a quarterly series, so it shows who is rising, who is sliding, and what moved the number.
The full methodology publishes with the Index: sampling design, brand resolution, scoring, and what we exclude. Reviewed by independent skeptics before release, and we publish what does not work, too.
Join the list to get Index releases, the methodology paper, and the executive brief on who is winning in AI answers.