Grounded
Answers cite their sources; retrieval over guesswork.
Auditable
Every run is logged and replayable, step by step.
Human-gated
Approvals and guardrails; agents escalate exactly when policy says.
Least privilege
Scoped access to systems and data — nothing more than the job needs.
Data handling & privacy
- ✓Encryption in transit and at rest; tenant isolation; configurable retention.
- ✓PII/PHI controls, redaction and consent capture; data-residency options (Enterprise).
- ✓Your data is not used to train shared models.
Access & governance
- ✓SSO/SAML, SCIM, RBAC, audit logs and exports.
- ✓Policy engine for guardrails, do-not-say lists and escalation rules.
AI safety & guardrails
- ✓Source-grounded retrieval to reduce hallucination.
- ✓Content controls and action approvals on sensitive steps.
- ✓Full transparency in the console — replay any decision.
Infrastructure & resilience
- ✓Cloud-hosted with tenant isolation, automated backups and defined disaster-recovery (RTO/RPO) targets — full details in the security overview.
- ✓Uptime SLA on Scale and Enterprise.
Built to meet GDPR · SOC 2 Type II · HIPAA/HITECH · ISO 27001 — certification in progress.
Security contact: security@perfox.ai · coordinated disclosure policy · bug bounty coming soon.