Disposition Engine

A disposition engine is the component of a runtime governance layer that decides what happens to each proposed agent action. It retrieves the controls that apply, evaluates the action against them, and returns one binding outcome before anything executes. The MAS SAFR (Safeguards for Agentic Finance at Runtime) white paper describes the output as "a defined, binding outcome for every proposed action calibrated to the specific risk it presents."

SAFR defines four outcomes. Deny: the action violates a hard regulatory or policy constraint, or its risk profile sits above defined thresholds; it is rejected before execution with a specific reason recorded. Escalate: the action is within scope and below hard constraints but above the threshold for autonomous execution; it is held pending human review. Auto-Execute: the action is within scope, below hard constraints, and within defined risk thresholds; it proceeds without human intervention. Observe: the action proceeds while a structured observation is logged for subsequent review, used where the institution has configured the action's pattern or signal as warranting attention.

SAFR identifies a pre-execution assurance gap: model risk management validates a system before deployment, and audit reviews what happened afterward, leaving the individual runtime decision unexamined. The disposition engine evaluates that decision at the moment the action is proposed. The engine evaluates each in-scope action deterministically against the retrieved controls, and every outcome is recorded with the specific rules applied and the basis for the result. Generic controls such as authorization checks and exposure limits are deterministic; AI-specific controls such as evidence quality may involve probabilistic or semantic assessment.

Outcome assignment is calibrated at design time on five factors: action reversibility, financial materiality, customer impact severity, regulatory sensitivity, and novelty relative to established patterns within the mandate. Higher-risk profiles shift outcomes toward Deny or Escalate.

In SAFR, the Disposition Engine is one of four runtime components. It works from the Controls Repository, evaluates each governance envelope, and writes every outcome to the Audit Log. In KLA's SAFR implementation, the KLA Policy Engine fills this role.