Association mapping systematically tests an AI model's parametric memory across three directional vectors: Entity-to-Brand (E2B) identifies which brands surface for category concepts, Brand-to-Entity (B2E) reveals what products or traits the model attributes to a brand, and Query-to-Brand (Q2B) captures which brands are recommended for specific search prompts. Analyzing these multi-round associative frequencies allows teams to quantify competitive share of voice, measure topic centrality, and pinpoint blind spots in an AI's unprompted recall.