Sandbox

Guided demo

The Sandbox is a live evaluation environment for how LSAS governs AI traffic before it reaches your model or downstream systems. Use guided scenarios or your own inputs to see deterministic decisions, redaction behavior, risk tiers, and audit-ready metadata on each request.

Key takeaway: LSAS turns policy into predictable runtime behavior your security, compliance, and product teams can review, explain, and operationalize.

Demo only. Do not send PHI, PCI, or production data. Requests may be logged for debugging.

1

Choose a scenario

2

Governance settings

Choose how strict LSAS should be when evaluating this request.

Policy tier
3

Provide input

Use text for chat-style prompts or a JSON payload for structured sanitize calls.

Input mode
Run a request to see LSAS decisions, risk tiers, and actions here.

Decision & outputs

Run a request to see an explanation.

Overrides

Awaiting baseline

Run a baseline request to evaluate governed override behavior.

Override execution is always explicit. Baseline requests run normally; an approved override is used only when you confirm and run the override action.

How to simulate an override

  1. Run a baseline request with Send request.
  2. If eligible, click Request override.
  3. Complete the override form, then click Submit governed override request.
  4. When approved, check the confirmation box and click Execute explicit override test.

Try these guided scenarios for eligibility

  • AI tool has access to protected data — Expected baseline: BLOCKED; use policy tier REGULATED or LOCKED_DOWN.
  • APIs exposing PCI data in transactions — Expected baseline: REDACTED; use policy tier REGULATED or LOCKED_DOWN.

Next step: run a baseline request first. Override controls activate only after a baseline decision exists.

Audit trail

Run a request to populate LSAS decision metadata.

Run a request to see the untrusted upstream payload.

Run a request and override workflow steps to populate authority, TTL, and post-review evidence.

The Sandbox provides a controlled evaluation environment for enterprise LSAS policy enforcement. Each execution returns a decision envelope with governance outcomes, risk classification, applied controls, and audit-ready metadata for review by security, compliance, and platform teams.

In production, LSAS integrates with your model providers, policy framework, identity controls, and observability stack to support routing, audit, and incident workflows. This environment is intentionally representative of real operating conditions so technical and executive stakeholders can validate control behavior before rollout.