Within the ARC framework, Relevance measures how consistently major AI platforms endorse and recommend a brand across specific entities, categories, and commercial search queries.

• Direct Evaluation Probing: Automated binary recommendation probes prompt models point-blank with structured queries across multiple test samples to calculate an empirical Query Relevance Score (QRS).
• Gap Detection: The system isolates low-scoring entity-brand pairs to pinpoint topics where competitors dominate the model's recommendation layer.
• Query Potential Weighting: By intersecting these probe scores with Google Search Console data, the system calculates Query Potential to flag high-volume search queries where weak model relevance represents the largest commercial opportunity.