Governance Intelligence

The Probabilistic Intelligence Plane — where ObsidianWall's advisory functions live, and how AI may advise without ever authoritatively governing.

Governance Intelligence is the other half of ObsidianWall’s Two-Plane Doctrine. It is where advisory functions live — analysis, correlation, explanation, and recommendation.

What This Plane Does

While the Deterministic Governance Plane decides whether a deployment is allowed, the Probabilistic Intelligence Plane asks a different set of questions:

Why does this policy keep getting overridden?
Which teams override most frequently, and why?
What is the financial exposure trend across
  governance decisions this quarter?
Where is governance drifting from intent?
Which controls are under-covering compliance
  obligations?

These are not yes/no questions with a single correct answer. They require pattern recognition across hundreds or thousands of governance decisions over time — exactly the kind of analysis probabilistic models are suited for, and exactly the kind of analysis that has no place inside an enforcement decision itself.

What Lives in This Plane

Compass                   Governance intelligence and economics
Governance Intelligence   Outcome correlation, pattern detection
Governance Economics      Financial exposure translation

The Rule

AI may advise. AI may explain. AI may optimize. AI may correlate. AI may recommend.

AI may NOT authoritatively govern.

Every output from this plane is advisory. Compass can tell a CISO that a specific policy generates overrides 40% of the time and suggest the policy may be miscalibrated. It cannot change the policy itself, and it cannot alter a Verdict decision that has already been made. The output of this plane informs human decisions — it never replaces them.

How This Closes the Feedback Loop

This plane is what makes the Assurance Feedback Loop possible. The Deterministic Governance Plane produces decisions. The Probabilistic Intelligence Plane observes patterns across those decisions and surfaces them back to the humans who author policy.

Verdict produces decisions
       ↓
Decisions accumulate in the Risk Ledger
       ↓
Compass analyzes patterns across decisions
       ↓
Compass surfaces insight to policy authors
       ↓
Policy authors refine intent
       ↓
Refined policy returns to the Deterministic
  Governance Plane
       ↓
[loop continues]

Without this plane, governance decisions would accumulate as a static record with no mechanism for organizational learning. With it, the pattern of past decisions actively informs future policy.

Forge — The Bridge Between Planes

Forge is the one component that lives in both planes simultaneously. It uses AI to help translate leadership intent into a draft governance policy — but that draft policy must pass through a human approval gate before it becomes an executable condition inside the Deterministic Governance Plane.

The human approval gate is the crossing point. AI can draft. AI can suggest. A human must approve before intent becomes enforcement.

Why the Separation Matters

If Governance Intelligence and Deterministic Governance lived in the same plane, the line between “AI suggested this” and “AI decided this” would blur the first time someone needed to explain a denied deployment to an auditor.

The Two-Plane Doctrine keeps that line explicit and structural, not just procedural. It is not a policy choice that could be overridden under pressure — it is the architecture itself.

Deterministic Governance → · Platform Architecture → · Vocabulary →