Compliance · EU AI Act

The EU AI Act, operationalised

Risk class and data sensitivity determine what the platform does — blocked, audited, labelled or routed as standard. Enforced on the control plane, recorded for the regulator.

Who it is for

Compliance & legal

Wants to demonstrate the EU AI Act to a regulator, not just describe it in a policy paper.

Data Protection Officer

Wants prohibited and high-risk use handled at the source, per class.

CISO & risk

Wants risk decisions deterministic and traceable, not dependent on a human judgement per case.

The problem

A policy document does not stop a prompt.

The EU AI Act sets requirements per risk class, but a PDF of guidelines enforces nothing. Without technical enforcement, "compliant" is a promise, not a fact.

Numaga translates the risk classes into behaviour on the control plane: prohibited use is blocked, high risk gets validation evidence, limited risk gets transparency labels — and every decision is immutably logged.

Risk-aware routing & data classification

The EU AI Act, operationalised.

Not as a guideline — as policy, enforced on the control plane.

Risk-class routing
EU AI Act classPlatform behaviour
Prohibitede.g. social scoringBlocked on the control plane.
High riskdecision & process supportAudit · validation evidence.
Limited risktransparency requiredTransparency labels added in-line.
Minimal riskStandard routing with monitoring.
Data-sensitivity routing
SensitivityBehaviour
PublicStandard routing.
InternalIdentity + audit.
ConfidentialEU-compliant provider + redaction.
Strictly confidentialpersonal data, trade secretsStrictest provider class + redaction + long-term retention.

Walk your AI Act obligations with us.

We show how risk classes are enforced and recorded on the control plane.

Get demo access