Seven gates decide whether an AI output should ever reach a decision-maker.
The envelope is uniform across every AI, agentic and Foresight iQ output on the platform. A gate either asserts, or it refuses to surface the output and records the refusal reason. Run the seven questions below against one of your own AI systems to see where the envelope would hold and where it would stop.
The check runs entirely in your browser. Nothing you enter is transmitted or stored.
G1 · Jurisdiction
Is the jurisdiction in which the output is consumed asserted and recorded?
G2 · Data residency
Do contributing data and computed output sit inside the agreed residency perimeter?
G3 · PII class
Is the PII or restricted-data class of every contributing input classified?
G4 · Model provenance
Are model version, training cut-off and provenance recorded with a model-card reference?
G5 · Drift posture
Are drift signals monitored against a baseline, with breach thresholds defined?
G6 · Human accountability
Is a named reviewer or reviewer queue assigned before the output is surfaced?
G7 · Audit and lineage
Does every output carry an audit identifier and replayable lineage?
Answer all seven gates to evaluate. Nothing is transmitted; the check runs in the browser.
A readiness answer is a declaration. Assurance requires the artefact behind it — which is what the inventory, the control library and the evidence vault exist to hold.