Skip to content

Security: Enterprise-ready security for your people data

People data is the most sensitive data a company holds. Every Zivo customer runs in its own environment, AI only reaches what each user is already allowed to see, and every change leaves a record. Certified ISO/IEC 27001:2022.

  • ISO/IEC 27001:2022 certification

    Certified by Prescient Security (certificate no. 122757, valid until March 2028). The scope covers our HR platform and its infrastructure on AWS, Vercel, Neon and GCP.

  • General Data Protection Law (LGPD)

    Data governance is led by a Data Protection Officer, as the LGPD requires, with recurring audits supported by the specialized law firm BZCP.

  • Independent penetration testing

    Manual penetration tests by Strike, a specialized security firm, on the same platform our customers use.

Zivo customers

How we protect your data

Four principles, built into the platform

Every control we run traces back to one of these principles.

  • Isolation per customer

    Each customer runs in a dedicated, single-tenant environment. Data integrations use a dedicated, read-only credential limited to your own company’s data.

  • Real identity

    You sign in through your own identity provider (SAML / OIDC). Blocking or offboarding someone in your IdP stops new access to Zivo. Production database access is restricted by role, with MFA and credentials rotated weekly.

  • Auditability

    Changes to business data are recorded in an audit trail with the responsible user, and all agent actions are audited too.

  • Defense in depth

    Data is encrypted in transit and at rest, with web access over HTTPS only. Permissions are checked on the server and in the database, never by the model.

AI under your rules

AI that never sees more than your people can.

Agents act with the permissions of the person who calls them, checked on our servers, and credentials are never exposed to the model. Your data, prompts and answers never train Zivo’s models or anyone else’s. Compensation recommendations come from deterministic rules you can trace: AI explains, the manager decides.

  • AI under your rules
  • Agents use the user’s access
  • No credentials exposed to AI
  • All agent actions audited
  • No training on your data
  • Your SSO (SAML / OIDC)
  • Your data, isolated
  • Single-tenant environments
  • Read-only data integrations
  • Encrypted in transit and at rest
  • Audit trail of changes
  • Restricted production access

For your security team

What security teams ask us first

Straight answers to the questions that come up in every enterprise review.

  • Where is our data hosted?

    On AWS, Vercel, Neon and GCP, all within the scope of our ISO/IEC 27001 certification.

  • Will our data train AI models?

    No. Our AI providers are contracted in no-training mode, so your data, prompts and answers never train Zivo’s models or third-party models.

  • Can we use our own identity provider?

    Yes. Sign-in runs through your IdP over SAML or OIDC, so your access policies and offboarding apply to Zivo automatically.

  • Has the platform been tested independently?

    Yes. Strike, a specialized security firm, runs manual penetration tests on the same platform our customers use.

  • What happens to our data when a contract ends?

    We return your data and issue a formal statement confirming its deletion.

Book a conversation with an
AI-native HR Transformation Specialist