Full definition
AI Search Ranking is the collective term for the signals and weights that generative search engines use to decide which sources to retrieve, paraphrase, and cite inside their generated answers. While each engine guards its exact algorithm, observable behavior across ChatGPT, Perplexity, Google AI Overviews, and Claude reveals a shared ranking model with five dominant signals. First, topical authority of the source domain, measured by the breadth and depth of topically-related content. Second, freshness — for queries with a temporal dimension, sources published or updated in the last 90 days are weighted higher. Third, content structure — pages with H2/H3 hierarchy, definition-first paragraphs, and FAQ schema are extracted more reliably. Fourth, entity clarity — pages that explicitly state the brand-entity relationship in the opening sentence. Fifth, off-site corroboration — the same fact appearing on three or more authoritative third-party domains. Harch Atelier models AI Search Ranking with a weighted scorecard per client and per query cluster, then runs weekly A/B content experiments to identify which signal moves the needle fastest for that specific brand. Ranking on AI search is not a single position; it is a probability of being cited, which rises or falls based on cumulative signal strength across all five dimensions.