Test a policy

Policy tests answer what Actera would decide for an exact input without creating an authoritative decision or performing the action.

Provide representative context

Choose a policy and edit the JSON input. Include the agent, action, resource, customer, and historical values the policy reads.

{
  "agent": { "id": "refund-agent" },
  "action": { "type": "refund.create", "attributes": { "amount": 80, "currency": "GBP" } },
  "customer": { "previous_refunds": 0, "fraud_flagged": false },
  "history": { "refund_total_24h": 1200 }
}

Read and compare the trace

Start with the final effect and resolution. Then inspect policies in groups: each rule shows whether its condition matched and what effect it would contribute. A matching lower-precedence rule remains visible even when a block determines the final result.

Comparison mode evaluates the current published bundle and draft against the same input. Focus first on outcome changes, then rules newly matched or no longer matched, reasons, and approval requirements.

Tests do not reserve usage capacity or prove that capacity will remain available during live, concurrent enforcement.