AI Decision Reproducibility and Testing

The answer

Corules decisions are reproducible by design. Each evaluation stores a policy set version and a normalized input hash. Given these two values, the exact evaluation can be reproduced at any future point — even after policy updates — by specifying the historical policy version. This enables: regression testing when policy changes (run historical cases against new policy), fairness testing (replay a sample of decisions across demographic segments), and regulatory examination (prove that the policy in effect at time T produced outcome O for input I).

How it works

Corules's policy runtime sits in the enforcement path between your AI agent and the action it wants to take. The agent sends a structured context payload to /v1/validate. Corules evaluates the context against a compiled CEL policy set and returns a structured decision — ALLOW, BLOCK, or ESCALATE — with a reason and audit ID.

Every decision is recorded in an immutable audit ledger. You can replay any past decision by providing the policy_set_version and the normalized input hash — the result will be identical.

Policy example

Policies are written in CEL (Common Expression Language). They are compiled once at publish time and evaluated in microseconds at request time.

// Replay any historical decision:
POST /v1/validate/replay
{
  "audit_id": "aud_01J...",
  "policy_set_version": "pset_v2.1.0"  // historical version
}
// Returns identical result as original evaluation

Frequently Asked Questions