ClaimGuardian AI
Governance dashboard
Trust surfaces expose the same backend services used by the agent pipeline: privacy gateway, decision ledger, evidence gate, fairness auditor, and state compliance.
PHI redaction stats
1,247
PHI entities redacted across 200 synthetic runs. 0 leaks detected.
Names86%
Member IDs72%
Dates of birth64%
Decision ledger
PlannerAgentplan_created8f31c0a9
PrivacyGatewayphi_redactedb944e177
EvidenceGatefinding_boundaac09125
DeepSeekCriticverdict_verified49ff8b12
Hash chain verified
Fairness metrics
Age-band DPD
0.028
Region DPD
0.034
Payer DPD
0.021
Rolling synthetic window stays under the 0.05 disparate-impact review threshold.
Responsible AI commitments
PHI redacted before any LLM call
Every claim is evidence-bound to a source
Human approval available at review points
All decisions logged to an immutable ledger
Self-consistency check on every critic verdict
Fairness metrics tracked across demographics