Governance & compliance

Enforce runtime policies, sanitize personally identifiable information (PII), and build toward EU AI Act compliance with Akka’s built-in governance capabilities. This pattern covers guardrails that validate agent inputs and outputs, sanitization pipelines that scrub sensitive data, and observability hooks for audit trails and compliance reporting.

Status: Partial — This pattern guide describes the governance approach and available components. A dedicated governance-focused sample project is pending; a full code walkthrough will be added when that sample is available.

Overview

When to Use This Pattern

  • You need to enforce content policies or safety guardrails on agent inputs and outputs

  • Your application must sanitize PII before it reaches an LLM or is stored in logs

  • You are building toward EU AI Act compliance and need audit trails for AI decisions

  • You want human-in-the-loop verification steps for high-stakes agent actions

Akka Components Involved

  • Guardrails — validate and filter agent inputs and outputs against defined policies

  • Sanitizers — detect and redact PII and sensitive data in agent interactions

  • Agents — integrate governance checks into the agent processing pipeline

Sample Projects

The following sample projects demonstrate aspects of this pattern:

A dedicated governance-focused sample may be needed for a full end-to-end walkthrough of this pattern.

See Also