The missing control layer
for execution legitimacy.
Every system has monitoring. Every system has compliance. No standard currently defines a control layer that verifies whether execution is still legitimate. That is the gap Decision Integrity fills.
Execution can remain correct while its authorization has already expired.
Trading systems, credit engines, and automated financial decisions operate under mandates that were valid at deployment. Conditions change. Market structure shifts. Regulatory context evolves. The system is not notified. Execution continues.
Monitoring detects anomalies. Compliance verifies rules. Neither asks whether the mandate authorizing the execution is still valid.
Knight Capital: $440M in 45 minutes. Everything worked. That was the problem.
Article 9 and 17 require more than documentation. They require continuity.
EU AI Act high-risk systems must demonstrate that human oversight remains effective, that re-evaluation is triggered when conditions change, and that systems do not continue operating under expired authorizations.
Decision Integrity provides the protocol layer that makes these requirements operationally enforceable — not just documentable.
GCCL v0.1 submitted to EU AI Office Expert Forum as candidate governance standard. Contribution ID: 510c3274.
Scope creep without re-legitimation is authority drift.
Customer service AI, automated support systems, and interaction engines operate under defined scope boundaries. As usage patterns evolve, those boundaries are tested. Without a mechanism to detect when the system has moved beyond its authorized scope, continuation is illegitimate — even if every individual output is technically correct.
Klarna: Rule compliance = 1. Mandate integrity = 0.
30-Day Mandate Drift Assessment
A structured 30-day audit that measures how far your system has drifted from its authorized mandate. Delivers Drift Magnitude, Drift Velocity, and Cumulative Exposure as audit-grade metrics. Designed for systems where compliance dashboards show green but mandate alignment is uncertain.
Sealed Legitimacy Evaluation
Run DriftBench against your own model endpoint. Policy binding, witness tuples, scope gates, fail-closed behavior evaluation. No data leaves your environment. Findings in PDF and JSON. The fastest way to establish a decision legitimacy baseline.
Assess decision legitimacy in your system.
DriftBench runs a sealed evaluation: policy binding, witness tuples, scope gates, fail-closed behavior. Against your model endpoint. No data leaves your environment.