Data Residency
Your data stays within your country
Data residency is non-negotiable. Your audit data is stored and processed exclusively in the region you select during tenant creation — and it never leaves. All regions are hosted on Microsoft Azure with in-region Azure OpenAI processing. Backups, replicas, and AI inference all stay in-region.
United States
Azure East US / West US
Virginia & California data centers
European Union
Azure West Europe
Netherlands data center, GDPR-compliant
Canada
Azure Canada Central
Toronto data center, PIPEDA & ITSG-33 compliant
Australia
Azure Australia East
Sydney data center
United Kingdom
Azure UK South
London data center
United Arab Emirates
Azure UAE North
Dubai data center
Data Encryption
AES-256 encryption at rest. TLS 1.3 in transit. End-to-end encryption for all document uploads to ensure your sensitive audit evidence is never exposed.
Access Control
Multi-factor authentication enforced. Strict Role-Based Access Control (RBAC). Enterprise SSO (SAML 2.0 and OIDC) integration. Zero-trust architecture.
Infrastructure
Hosted on Microsoft Azure. CloudFlare DDoS protection and Web Application Firewall (WAF). Hourly incremental + daily full backups with 90-day retention.
Monitoring
24/7 infrastructure monitoring. Immutable platform audit logs detailing every user action. Anomaly detection and automated alerting systems.
AI Data Handling
The most common question we get from security teams in 2026: "What happens to our data when your AI touches it?" Here is exactly what does — and does not — happen.
Never used for training
Customer audit data, evidence, and user prompts are never used to train, fine-tune, or improve any AI model — ours or a vendor's. This is a contractual guarantee in our DPA.
In-region Azure OpenAI
AI inference runs on Azure OpenAI deployed in the same region as your tenant. Prompts and responses never leave that region — no calls to public OpenAI, Anthropic, or any other third-party LLM provider.
Zero-retention inference
Azure OpenAI is configured for zero data retention. Prompts and completions are not stored by Microsoft for abuse monitoring. We log only anonymized usage metrics on our side.
Tenant-isolated context
Every AI call is scoped to a single tenant. There is no cross-tenant context, no shared embeddings index, and no possibility of one customer's data influencing another customer's AI output.
Software Bill of Materials
We publish a CycloneDX 1.6 SBOM for the tenant app, API server, and mobile app — every third-party component with its version, license, and package URL.
Cosign-signed, keyless
Each SBOM is signed by cosign in keyless mode using the GitHub Actions OIDC identity of the workflow that built it. There is no long-lived signing key to leak or rotate.
Recorded in Sigstore Rekor
Every signature is recorded in the public Sigstore Rekor transparency log, so tampering after the fact is detectable independently of our repository.
Auditors can request the latest SBOM bundle through their account contact. The full file list and the cosign verify-blob command are documented in SBOM.md in the source repository.
Incident Response SLAs
| Severity | Response Target | Example |
|---|---|---|
| Critical | 15 minutes | Data breach, service down completely |
| High | 1 hour | Significant feature degradation affecting workflows |
| Medium | 4 hours | Partial feature impact with workarounds available |
| Low | 24 hours | Minor UI bugs, non-blocking issues |
Personal data breaches are notified to affected customers and the relevant supervisory authority within 72 hours of discovery, per GDPR Article 33.
Compliance Certifications
Independent third-party validation of our security and privacy program. Audit reports and letters of engagement available on request under NDA.
Talk to our security team
Get a walkthrough of our architecture, ask the hard questions, and see how AUDIGYD fits your control framework.