Security & compliance
Last updated:
Our Security Philosophy
We treat your data and agent memory as our highest responsibility. Every system we build is private, encrypted, and auditable by design.
Core Security Controls
Encryption: Data encrypted at rest (AES-256) and in transit (TLS 1.3)
Private Deployments: Client-isolated environments (no shared infrastructure)
Access Controls: Role-based access, least-privilege principle, SSO/MFA support
Audit Logging: Full immutable logs of every data access and agent interaction
Versioning & Recovery: Complete history and rollback capabilities
Threat Protection: Continuous monitoring, anomaly detection, and regular penetration testing
Compliance Certifications & Standards
SOC 2 Type II (Security, Availability, Confidentiality, Processing Integrity)
ISO 42001 (AI Management System) ready
GDPR & CCPA compliant by design
Zero data retention for model training without explicit consent
How We Protect Your Agents
Memory systems are never used to train external models
Intelligent data classification and redaction tools
Secure multi-agent memory sharing with granular permissions
Incident Response We maintain a formal incident response plan with 24-hour notification for any breach affecting client data.
Transparency All clients receive a Security & Compliance Pack upon project kickoff, including architecture diagrams and audit rights.
Questions? security@automat.ai We are happy to provide custom security questionnaires or join your procurement review.
