Trust & data handling

This page summarises how Buzzi.ai handles data gathered by the Shadow AI Audit. It is deliberately specific so that your CISO, DPO, and — in the case of the Google Workspace scan — Google CASA reviewers can confirm our posture.

Google Workspace scan (ephemeral)

When a Workspace admin clicks Scan Google Workspace, we request two read-only scopes on an incremental consent screen:

  • admin.directory.user.security.readonly
  • admin.reports.audit.readonly

We enumerate the OAuth apps granted by users in the domain and match them to our tool registry. We store the aggregated result (tool name, client ID, user count, first-granted timestamp) in workspace_scan_sessions. Nothing is sent to any third-party model, analytics, or CRM.

Retention: all scan data is automatically purged within 24 hours. The expires_at column is enforced by a daily cron (/api/tools/shadow-ai/cron/scan-purge). You can also trigger deletion at any time by contacting privacy@buzzi.ai.

Audit submission data

The tools you identify, your survey answers, and the contact details you enter are stored in our Supabase Postgres instance. We use this only to produce your report and — where you opted in — to contact you with Buzzi.ai insights.

  • Hosting: Supabase Postgres (AWS, US / EU regions).
  • Backups: encrypted at rest; retained per vendor defaults.
  • Retention: audits kept for 24 months; email stored for unsubscribe purposes indefinitely unless deletion is requested.
  • Sub-processors: Supabase (hosting), Postmark / Amazon SES (email), OpenAI and Google (LLM outputs).

LLM-generated artefacts

When you generate the executive summary, DPIA, or vendor questionnaire, we send the audit context to our contracted LLM providers (OpenAI and Google) under enterprise agreements that exclude training on our customer data. Outputs are stored in audit_outputs scoped to your account.

Your rights

You can request a copy, correction, or deletion of your audit data at any time by emailing privacy@buzzi.ai. For EU residents, this is your GDPR Article 15 / 17 request.

Security incident response

If you believe you have found a security issue, please disclose responsibly via security@buzzi.ai. We will acknowledge within 2 business days and follow our standard IR procedure including, where appropriate, regulator and data-subject notification.