Skip to content

Security

How we protect your workspace and evidence

AI governance work holds sensitive facts about your systems, suppliers and decisions. This page describes the controls in place today, in plain terms, so you can check them against your own requirements.

Access

Signing in and who sees what

  • Sign in with Google or Microsoft, or with a short-lived, single-use email code. Stored codes are hashed and sign-in attempts are rate-limited.
  • Every organisation has its own workspace. Each page, record and file is checked against the signed-in person's organisation.
  • Members are Owners, Administrators, Contributors or Viewers, and only Owners and Administrators approve decisions.
  • Suppliers and independent specialists use secure single-use invitation links that give access to their task only.

Evidence

How evidence files are handled

  • Managed evidence is held in private, encrypted Amazon Web Services object storage.
  • Files upload straight to storage through a link that expires after five minutes, and each file's SHA-256 checksum is verified.
  • Each file is scanned for malware before anyone can download it. A file that fails the scan is quarantined.
  • Downloads are authorised per organisation and recorded. Or keep evidence in your own system and store only a reference to it.

Records

A trail you can review

  • Decisions, approvals, evidence access and administrative changes are written to audit logs.
  • Reports and documents carry a version and a content hash, so a later copy can be matched to the original.
  • Managed storage tiers include retention review and controlled deletion.

The website

Browser and connection security

  • Every page and the workspace use encrypted HTTPS connections.
  • A strict content security policy limits which scripts, frames and form destinations the browser will accept.
  • Signed-in pages run only scripts that carry a new one-time code for each request.

Your data

What we do and do not do with it

  • We do not sell your data, and we do not use it to train AI models.
  • Results come from published rules that a person reviews. Workspace content is not sent to an AI model.
  • The free exposure check works out its result in your browser. We keep only the result category and the time.
  • Optional website analytics runs only if you accept it.

Assurance

Certification and questionnaires

Certification

EU AI Fit does not yet hold an independent security certification such as ISO 27001 or SOC 2. We describe the controls that are in place instead of claiming a badge we have not earned.

Hosting, subprocessors and questionnaires

Current hosting locations, subprocessors, retention terms and transfer arrangements are shared during onboarding. We also answer customer security questionnaires. See the privacy policy for how personal data is handled.

Contact

Report a security concern

If you think you have found a security issue, email support@euaifit.com with “Security concern” in the subject. Please do not include evidence files or personal data. We will reply and keep you informed while we investigate.