Direct relevance testing prompts models with binary 'yes/no' questions across repeated sampling runs to measure whether an AI considers a specific brand authoritative for a given query or entity, producing a quantifiable QRS metric.