Deterministic Core — Passenger Journey
REAL logic · source of truthAI never decides anything here. Each step is a deterministic rule, fully audited. Biometric template & matching are simulated.
Try the RBAC: switch the role to Security Admin or System Admin and click
“Read full identity” — the server will deny it.
Admins can run the system but cannot read personal data.
AI Assistive Layer — 2 safe features
advises · never decides
(Security Admin only · core keeps running either way)
1 · NL Reporting Helper
Aggregated, non-personal data only (🟢). No human review needed.
2 · Audit-Log Triage Helper
Masked/tokenized data (🟠). Human review required.
Auditor / Security Officer / Privacy Officer
Every AI call passes through the policy gateway (kill switch → RBAC → masking → AI audit → output label).
Outputs are always labeled “AI-generated suggestion — verify before acting.”
Integration Layer — simulated connectors (Stage 6)
token-based · auditedEach connector is simulated. Toggle one DOWN (System Admin), then run the journey to see fail-safe fallback. The on-prem matching engine (INT-6) keeps gates working even if external doors fail.
Live request/response flow
🔒 on-prem · 🌐 external boundaryRun the journey to see connector traffic…
Live state (non-personal)
loading…
Response
Click any action…