Trust & Security

Netallion AI Assurance is built with enterprise security requirements at its foundation. We protect your data with the same rigor we help you apply to your own infrastructure.

Compliance Alignment

SOC 2 Aligned

Controls mapped to SOC 2 Type II trust service criteria. Audit evidence exportable from the platform.

HIPAA Aligned

PII/PHI detection, tamper-evident audit trails, and role-based access controls meet HIPAA security requirements.

PCI-DSS Aligned

Credit card number detection, encryption controls, and access logging aligned with PCI-DSS requirements.

GDPR Compatible

Data minimization (no raw secrets stored), comprehensive audit logs, and complete tenant isolation.

SOC 2 Type II Audit In Progress

Observation period starting June 2026. Report expected October 2026.

"Aligned" means Netallion AI Assurance implements controls mapped to the referenced framework. This is not a certification or attestation. See CONTROL_ALIGNMENT.md for full details.

Security Architecture

Encryption at Rest

All stored data encrypted with Fernet symmetric encryption. Database-level encryption for all tenant data.

Encryption in Transit

TLS 1.2+ enforced for all API and data connections. No plaintext data transmission.

Row-Level Security

PostgreSQL row-level security policies ensure complete tenant data separation at the database layer.

Zero Raw Secrets

Netallion AI Assurance never stores raw secret values. Only redacted representations, metadata, and finding details are persisted.

Tamper-Evident Audit

Hash-chain audit logging ensures complete integrity of all security records. Every action is recorded and verifiable.

Token Security

Redis-backed token blacklisting with TTL expiry. Single-use OIDC transaction store prevents replay attacks.

Data Flow Architecture

  Your Infrastructure                  Netallion AI Assurance Platform
  ===========================         ===========================

  Azure Monitor Workspaces              Detection Engine
  [Logs] ─────────────────────────────> [467 Patterns + BPE]
                                              │
  GitHub / GitLab Repos                       │
  [PRs / MRs] ───────────────────────> [Live Verification]
                                              │
  Collaboration Tools                         │
  [Slack, Teams, Jira] ──────────────> [Finding Classification]
                                              │
  AI Prompts                                  │
  [ChatGPT, Copilot Chat] ──────────> [Policy Evaluation]
                                              │
                                              v
                                     ┌─────────────────┐
                                     │ Encrypted Store  │
                                     │ (Redacted Only)  │
                                     │ RLS per Tenant   │
                                     └────────┬────────┘
                                              │
                                     ┌────────v────────┐
                                     │  Dashboard &     │
                                     │  Remediation     │
                                     │  (TLS 1.2+)      │
                                     └─────────────────┘

Data Handling

What we store

  • Redacted secret representations (first/last 4 characters only)
  • Finding metadata: type, location, severity, verification status
  • Scan results and incident records
  • User actions and audit trail entries
  • Configuration and policy settings

What we never store

  • Raw secret values (keys, tokens, passwords)
  • Full log contents from Azure Monitor
  • Source code from repositories
  • Full text of collaboration messages
  • AI prompt content (only scan results)

Retention policies

  • Starter: 30-day finding retention, 30-day audit logs
  • Professional: 180-day finding retention, 90-day audit logs
  • Enterprise: Configurable retention, unlimited audit logs with hash chain
  • All data permanently deleted upon tenant removal request
  • Backups follow the same retention schedule as primary data

Have security questions?

Our security team is available to discuss architecture, compliance, and data handling in detail.