Preview Release: Early-access preview, information and pricing subject to change.

The AGEI Governance Framework

A comprehensive model for governing autonomous AI agents based on capability, authority, admissibility, and evidence.

Core Principles

Capability does not imply authorization.

Just because an agent can perform an action doesn't mean it should be allowed to.

Authorization does not imply admissibility.

Permission to act requires appropriate context, risk assessment, and oversight.

Execution requires governance.

All agent actions must be evaluated, decided, overseen, and evidenced.

Governance Lifecycle

Agent

Policy

Connector Event

Decision

Human Oversight

Evidence

Six Governance Dimensions

Authority

What actions can the agent perform? From read-only to admin-level execution authority.

Context Access

What data can the agent access? From public to regulated and critical operational data.

Memory Scope

What information can the agent retain? From stateless to persistent cross-system memory.

Autonomy

How much human oversight is required? Human-in-loop, on-loop, or out-of-loop.

Human Oversight

Who reviews high-risk actions? HITL requires approval, HOTL monitors, HOOTL operates autonomously.

Impact

What are the consequences of failure? From low to critical financial, regulatory, or operational impact.

Four Governance Tiers

Automatically calculated based on the six governance dimensions

Tier 1: Limited

Informational assistance, low impact, human-in-the-loop required

Tier 2: Controlled

Defined workflows, limited authority, active oversight and approval required

Tier 3: Material

Tool use, execution authority, sensitive data access, high impact potential

Tier 4: Critical

Open-ended operation, broad authority, critical impact, maximum oversight

Nine Agent Types

Classify agents by primary function to determine appropriate governance

Assistant Agent

Workflow Agent

Routing Agent

Monitoring Agent

Decision Support

Tool-Using Agent

Execution Agent

Orchestration Agent

Strategic Agent

Implement Evidence-Based Governance

Start governing your AI agents with the AGEI framework