{
  "schemaVersion": 1,
  "generatedAt": "2026-08-20",
  "totalInsights": 3,
  "methodology": "Derived from evidence-backed broker coverage records captured through 2026-08-20. These are descriptive cohort observations, not universal industry claims or removal-success rates.",
  "insights": [
    {
      "id": "insight-route-evidence-coverage",
      "title": "Route Reachability and Request Outcomes Are Separate Evidence States",
      "category": "FRICTION_PATTERNS",
      "observation": "761 read-only route checks are recorded across 1018 catalog entities; no request outcome is inferred from route reachability.",
      "sampleSize": 1018,
      "implication": "Consumers should retain submission and response evidence separately from a broker page check."
    },
    {
      "id": "insight-email-workflow-observations",
      "title": "Email Support Is Counted Only When the Workflow Source Documents It",
      "category": "DATA_PRACTICE_TRENDS",
      "observation": "952 records have a source-backed email workflow observation; 471 records remain unknown for government-ID requirements.",
      "sampleSize": 1018,
      "implication": "The benchmark reports documented observations rather than guessing from broker category or contact fields."
    },
    {
      "id": "insight-registry-crosswalk-scope",
      "title": "Registry Crosswalks Are Snapshot-Bounded",
      "category": "REGULATORY_ENFORCEMENT",
      "observation": "817 entities have normalized registry profiles in the captured state snapshots; absence from a captured snapshot is not proof that an entity is outside every registry.",
      "sampleSize": 1018,
      "implication": "Use the jurisdiction and capture date with every registry comparison."
    }
  ]
}