Skip to main content

Decision Engine

The Decision Engine turns Trust Graph context + Policy Engine rules into a reproducible verdict.

What you get back

  • Verdict (approve / review / decline — program-defined)
  • Recommended limit or access tier
  • Principal reasons (reason codes)
  • Model version and policy version
  • Conditions and monitoring hooks

Design rules

  • Same inputs + same policy version → same path
  • Sandbox uses deterministic rules for demos and CI
  • Live / production keys run decisions under your program’s compliance posture when Vence enables live for your org (live financial underwriting is gated)
  • AI assists investigation; synchronous auth prefers deterministic policy
Call underwrite via the API overview or follow the underwrite guide.