Skip to content

Security and Privacy

Aliveo AI is built with security and privacy as foundational principles, ensuring that customer data is protected at every layer of the architecture. Our approach includes strict data isolation, encryption, identity controls, and continuous monitoring to maintain enterprise-grade security.

Compliance Posture

  • SOC 2 Type II controls cover access management, change management, data handling, vendor risk, and incident response.
  • Privacy-by-default: customer data is only ever used to operate the service for that customer. We do not train shared models on customer data.
  • AI sub-processors: to answer your questions, Aliveo sends the relevant parts of your questions and data to large language model providers acting as sub-processors: Microsoft Azure OpenAI, OpenAI, Perplexity (live web search) and AWS Bedrock (creative analysis).
  • Customer-specific compliance addenda (GDPR, HIPAA, regional residency) are available for enterprise plans—reach out to your Aliveo contact.

Architecture and Network Isolation

Our platform defaults to a multi-tenant model but can also run single-tenant on request:

  • Multi-Tenant (Default): Customers share a common VPC with logically isolated subnets (front-end APIs, application services, databases) enforced by strict NACLs and security groups. This delivers lower cost, faster onboarding, and centralized updates.

  • Single-Tenant: Each customer gets its own VPC with separate subnets for front-end APIs, application services, and databases. Dedicated resources provide stronger isolation, no “noisy neighbors,” and full control over network policies—ideal for strict compliance or custom configurations.

Storage and Encryption at Rest

Persistent data resides in individual database instances or object storage buckets dedicated to each tenant. Ephemeral storage, such as cache volumes, also exists only within the customer’s VPC. All storage resources are encrypted using customer-managed keys (CMKs) via a Key Management Service (KMS), ensuring that data, backups, and snapshots remain encrypted. Key rotation is handled by the customer’s KMS policy, and Aliveo AI’s monitoring tools verify active key usage.

Encryption In Transit

External client-to-API communication uses HTTPS with a minimum of TLS 1.2, while internal service-to-service traffic is routed through a service mesh enforcing mutual TLS (mTLS). API gateways, ingress proxies, and database drivers are configured to require encrypted connections. If customers opt for private connectivity, network-layer encryption (IPSec or MACSec) provides an additional layer of protection.

Processing Layer and Access Controls

Container images undergo vulnerability scanning before deployment, and secrets are injected via a cloud provider’s encrypted secrets management service. Authenticated APIs enforce tenant-scoped requests using JSON Web Tokens (JWTs), ensuring that data operations are limited to the correct customer. IAM roles and service accounts follow the principle of least privilege. Access to production systems is restricted to authorized Aliveo engineers and is used only to support your account, investigate issues and operate the service.

Identity, Authentication, and Authorization

Authentication and authorization protect both the platform UI and its APIs.

Sign-in methods

  • Email + password with strength requirements and breached-password rejection.
  • TOTP-based two-factor authentication (2FA) for any user from their personal account settings.
  • SSO/SAML for enterprise customers—identity provider becomes the source of truth for membership.

Organization-wide 2FA enforcement

Primary account holders can require 2FA for every member of the team. Once enforced, users who haven’t yet enabled TOTP are routed to a Set up Two-Factor Authentication screen on their next login and cannot access the app until enrollment is complete.

Step-up authentication for sensitive actions

Even after you’ve signed in, some actions require a fresh email one-time code to proceed. These include:

  • Changing your password.
  • Inviting, enabling, or disabling team members.
  • Toggling 2FA enforcement.
  • Changing Slack permissions (allowing non-Aliveo users to call the bot, or allowing actions to be applied from Slack).
  • Turning AI client (MCP) access on or off.
  • Changing a team member's actioning permissions.
  • Regenerating the SCIM token.

This is a defense-in-depth measure: if a session token is somehow compromised, the attacker still can’t escalate without access to the user’s email.

Roles

Aliveo has two user roles:

  • Primary — administrators who can manage integrations, context, team members, and security settings.
  • Secondary — invited teammates who use the platform with scoped access.

Either role can also be set up as read-only, which allows viewing but not creating or changing anything. See User Management.

Internal employee accounts (@aliveo.ai) are hidden from customer admin views unless the primary user is themselves an Aliveo employee.

Actioning permissions

Pushing changes back to ad platforms ("actioning") is governed by a per-user permission layer, separate from read access:

  • Primary users can action by default.
  • Secondary users can action only for the accounts or platforms they've been explicitly granted, per organization.
  • Read-only users can never action.

These permissions are enforced on the server for every apply request, so what a user sees in the UI can't be used to bypass them. See Actioning Permissions.

Security notifications

Sensitive account changes trigger a security notification email to the affected user and the account's primary user. These include adding, disabling, or re-enabling team members, password and 2FA changes, turning org-wide 2FA enforcement on or off, regenerating the SCIM token, and turning Slack actioning or AI client (MCP) access on or off.

Every change applied to an ad platform through Aliveo is also recorded in the Action Log — what changed, its before → after values, who applied it, and where it came from (a chat, a workflow, or a restore) — and most changes can be restored with one click. See Action Log & Restore.

Data Lifecycle and Retention

  • Connected ad-platform data is fetched via OAuth, or an API key for platforms such as Moloco. The long-lived OAuth refresh tokens and API keys this needs are stored encrypted at the application level with AES-GCM. Your ad-platform passwords are never stored.
  • Custom data (CSV, Sheets) is stored in your tenant’s encrypted buckets and can be deleted on demand.
  • On account closure, customer data is purged from production storage within 30 days; cold backups are purged on the standard backup retention rollover.

Observability and Incident Response

Aliveo monitors agent runs and platform health in real time:

  • Structured tracing captures every planning, code-execution, and tool-call step (see Evaluations and Observability).
  • Anomaly detection alerts on schema drift, null spikes, and elevated error rates.
  • A documented incident-response runbook covers detection, containment, customer notification, and post-mortem reporting.

Customers who require additional contractual notification timelines or dedicated escalation paths can configure them via their enterprise agreement.


For the user-facing security controls (2FA, OTP reauth, disable/enable, SSO), see User Management.