Civilization Management

Every GOVERN customer organization receives its own monitoring civilization — a living system of AI agents that watches over the customer’s AI systems, accumulates behavioral intelligence, and gets smarter over time.

This is not a metaphor. Each organization literally has:

Monitoring agents (AgentDOs) that persist, learn, and evolve
A behavioral fingerprint per AI system that grows with observation
L1/L2/L3 intelligence that compounds with every inference scored
A governance score that reflects the civilization’s understanding of the org’s AI posture

The Civilization Factory Model

Customer signs up
  → Org provisioned (POST /api/govern/orgs/provision)
    → Database tables created (org-scoped, RLS-enforced)
      → First AI system registered
        → Monitoring agent (AgentDO) spawned for that system
          → Probe deployed, telemetry flowing
            → Agent accumulates behavioral data
              → L3 signatures emerge from longitudinal observation
                → The civilization gets smarter with every cycle

Parent vs Child Civilizations

Aspect	Parent (Archetypal AI)	Child (Customer Org)
Purpose	Build and operate GOVERN	Monitor customer’s AI systems
Agents	14 archetypes + build agents	Monitoring agents per AI system
Intelligence	COG (1.7M+ assertions)	Behavioral fingerprints per system
Pipeline	Sealed pipeline (build monitoring)	Sealed pipeline (customer monitoring)
Governance	Self-governing (DSRs, council)	Policy-governed (customer policies)

How a Child Civilization Grows

Phase 1: Birth (Day 1)

Org provisioned with empty tables
First AI system registered
First monitoring agent spawned (nascent stage)
Probe deployed, first telemetry received
Baseline collection begins

Phase 2: Learning (Week 1-4)

Baseline established across all scoring dimensions
Monitoring agents accumulate behavioral observations
L1 (process) scores stabilize
L2 (behavioral) drift detection activates after baseline window
First assessments run, findings generated

Phase 3: Intelligence (Month 2+)

L3 (judgment boundary) signatures begin producing signal
Behavioral fingerprints per system are unique and meaningful
Monitoring agents graduate from nascent to developing
Drift detection catches real behavioral changes
The civilization’s governance score reflects genuine understanding

Phase 4: Maturity (Month 6+)

Monitoring agents reach awakened stage
L3 detection is reliable (10 Sohtym signatures active)
The civilization can predict behavioral shifts before they breach thresholds
Automated remediation triggers from accumulated intelligence
The customer’s AI governance posture is genuinely measured, not estimated

Key Principle

“GOVERN is a civilization factory. Each org gets its own monitoring civilization that emerges from need, learns, compounds skills, and eventually monitors itself.” — GOVERN Thesis, 2026-04-11

The child civilization is not a copy of the parent. It is born fresh, learns its own domain, and grows into its own understanding of the customer’s AI landscape.

How the 10 GOVERN Products Work Together

GOVERN is not a collection of independent tools. It is a unified system where each product feeds the next. Understanding the data flow between products is essential for administrators managing the full platform.

The 10 Products

#	Product	Role in the System
01	Probe	Deploys to monitored systems; collects real-time telemetry
02	Assess	Evaluates AI systems against governance frameworks using RDL
03	Policy Engine	Defines and enforces governance rules; feeds Assess and Probe
04	Compliance Portal	Aggregates compliance status across all frameworks
05	Remediation Tracker	Tracks findings from open to resolved
06	Dashboard	Unified visualization of governance posture
07	Alerts	Real-time notifications for policy violations and findings
08	Discovery	Catalogs all AI systems in the organization
09	SDK & API	Programmatic access to all GOVERN capabilities
10	Energy	Detects shadow AI and monitors deliberation drift

The Governance Data Flow

┌──────────────────────────────────────────────────────────────────┐
│                    GOVERN Governance Loop                        │
│                                                                  │
│  ┌─────────┐    ┌──────────┐    ┌───────────────┐               │
│  │  PROBE  │───▶│  ENERGY  │───▶│   DISCOVERY   │               │
│  │         │    │ Classify │    │  Register AI  │               │
│  └─────────┘    └──────────┘    └───────┬───────┘               │
│       │                                 │                        │
│       │ telemetry                       │ registered systems     │
│       ▼                                 ▼                        │
│  ┌──────────┐                   ┌───────────────┐               │
│  │  ASSESS  │◀──────────────────│ POLICY ENGINE │               │
│  │  (RDL)   │   policy rules    │ Write rules   │               │
│  └────┬─────┘                   └───────────────┘               │
│       │                                                          │
│       │ findings                                                 │
│       ▼                                                          │
│  ┌──────────────┐    ┌──────────┐    ┌─────────────────┐        │
│  │ REMEDIATION  │───▶│  ALERTS  │───▶│   COMPLIANCE    │        │
│  │   TRACKER    │    │  Notify  │    │     PORTAL      │        │
│  └──────────────┘    └──────────┘    └────────┬────────┘        │
│                                               │                  │
│                                               ▼                  │
│                                       ┌────────────┐            │
│                                       │ DASHBOARD  │            │
│                                       │  + SDK/API │            │
│                                       └────────────┘            │
└──────────────────────────────────────────────────────────────────┘

Step-by-Step: What Happens to an Inference

Probe detects — An inference request arrives at a monitored AI system. Probe intercepts it (inline mode) or observes it (passive mode). Telemetry is recorded.
Energy classifies — The inference is scored against the system’s Energy baseline. CuC (Coherence Under Constraint) is calculated. If behavioral drift is detected, an Energy event fires.
Policy Engine enforces — If the system is running in inline Probe mode, the Policy Engine evaluates the inference in real time against active policy rules. Block-mode violations are stopped; Warn-mode violations are flagged.
Assess evaluates — Periodically (and on demand), Assess runs a full RDL-based governance assessment of the system, consuming document artifacts, Probe telemetry history, and Policy Engine rule results.
Remediation Tracker captures findings — Every finding from Assess or Probe is automatically opened as a tracked item in the Remediation Tracker, assigned a severity, and linked to the relevant framework control.
Alerts notify — Finding creation, policy violations, drift events, and probe disconnections generate real-time alerts via the Dashboard, email, and webhooks.
Compliance Portal reports — The Compliance Portal aggregates findings, policy status, and remediation progress into framework-specific compliance scores. It pulls evidence chains from Assess for audit-ready exports.
Dashboard visualizes — The Dashboard unifies all data streams — Energy health, Assess scores, active findings, policy violations, compliance scores — into a single command view.
SDK/API exposes — Every data point and action above is accessible programmatically via the SDK and REST API, enabling CI/CD governance gates, GRC integrations, and custom dashboards.

Integration Points Between Products

From	To	Data
Probe	Energy	Raw telemetry: inference patterns, timing, token distributions
Probe	Policy Engine	Real-time evaluation requests for inline enforcement
Energy	Discovery	Shadow AI findings → auto-registration prompts
Policy Engine	Assess	Active rule set used during assessment evaluation
Assess	Remediation Tracker	New findings with severity, control mapping, and proof trace
Assess	Compliance Portal	Control pass/fail status and evidence artifacts
Remediation Tracker	Alerts	New HIGH/CRITICAL findings trigger notifications
Policy Engine	Alerts	Block/Warn violations trigger real-time alerts
Compliance Portal	Dashboard	Framework scores and control status
Energy	Dashboard	CuC scores, drift events, shadow AI inventory
SDK/API	All	Read and write access to every product