Skip to content
Devernance

Product

An end-to-end assurance platform for AI features.

Turn regulatory requirements into practical, auditable actions. Connect risk, evidence, evaluations, data contracts and approvals into a single release decision.

Sample data

Claims Triage AI

High risk
Search evidence
OverviewEvidenceEvaluationsContractsApprovalsEvidence pack

Release readiness

  • Risk classificationHigh risk
  • Evidence coverage5/5
  • Evaluation results0.92
  • Data contractsHealthy
  • Human approvals3/3

Release decision

PASS

All mandatory controls satisfied

Export signed evidence pack

Inside the platform

Every signal a release decision needs.

Evidence, evaluations, data contracts and release gates have their own pages. These are the pieces that connect them.

AI System Registry

The register every later step reads from: owner, purpose, risk class, deployment context, vendor and model, data sources, affected users and release status.

  • Risk class with rationale
  • Deployment context and data sources
  • History of every change

Risk & Control Mapping

Systems are classified against a pinned EU AI Act corpus. Control mappings arrive as proposals; a person accepts them before they count.

  • Pinned legal corpus with EUR-Lex citations
  • Proposal queue, never auto-applied
  • Control modes: informational to blocking

Approvals

REVIEW and BLOCKED systems route through engineering, compliance and legal sign-off. Reviewer identity comes from the session, and every decision lands in a hash-chained ledger.

  • Multi-stage sign-off
  • Append-only audit ledger
  • Ledger verification endpoint

Audit & Evidence Packs

Export the evidence behind a release decision as canonical JSON plus PDF, signed with RS256 and verifiable against the public JWKS.

  • Signed, verifiable export
  • Annex IV-shaped checklist
  • Same gaps as the open-source evgraph scan

Before AI ships, prove it is ready.

Free plan, no card. Gate your first AI system in CI today, or explore the live demo workspace.