Understanding Leak Findings

How to read leak details — severity, impact, confidence, and recommended actions.

Finding Cards

Each leak finding shows:

  • Title — a brief description of the leak
  • Severity — Critical, High, Medium, or Low
  • Estimated Impact — dollar amount of revenue at risk annually
  • Confidence Score — how certain the AI is about this finding (0-100%)
  • Category — which of the 12 leak categories it belongs to
  • Recommendation — specific steps to fix the leak

Actions You Can Take

  • Fix — mark a finding as fixed. On the next scan, LeakShield will verify the fix
  • Dismiss — mark as not relevant. The AI learns from dismissals and adjusts future analysis

Confidence Score

The Validator agent assigns confidence scores based on data quality, evidence strength, and cross-validation. Scores above 75% are considered high-confidence. Lower scores may need manual review.

Ready to Find Your Revenue Leaks?

LeakShield AI uses autonomous agents to monitor your revenue streams 24/7 — detecting leaks, prioritizing by impact, and showing you exactly how to fix them.

Get Started