Loading AGEI...
How AGEI governance evidence infrastructure can support ADG-style AI control evidence, audit review, and operational governance.
The EC-Council Adopt. Defend. Govern. framework describes enterprise AI governance across pillars, surfaces, minimum controls, operating artifacts, autonomy tiers, and deployment overlays. AGEI is not a replacement for ADG, NIST AI RMF, ISO/IEC 42001, or legal compliance review. AGEI helps create the operational evidence trail that governance teams, auditors, risk teams, and technical teams need to review.
ADG defines the governance expectations.
AGEI helps create the evidence trail.
AGEI provides strong current support for these ADG-aligned evidence areas
Core architecture principle
AI systems cannot approve themselves. Server-side gates enforce approve/deny/escalate/inspect decisions with HITL escalation when needed.
Complete data capture
AGEI-controlled workflows generate cryptographic receipts for governed events.
Strong foundation
Gate system validates training quality, deployment readiness, and policy compliance before release.
The following table compares the 12 ADG Minimum Controls with AGEI's current capabilities and support status.
Support levels are based on current AGEI product capabilities and may require configuration, templates, or design-partner implementation work.
ADG Expects: Inventory with owner, risk classification, lifecycle state, and autonomy tier.
AGEI Support Today:
AI lifecycle object registry, lifecycle links, receipts, evidence object references.
Current Status:
Strong schema support; dedicated inventory UI and ADG profile export needed.
ADG Expects: Risk tier based on harm, autonomy, data sensitivity, and deployment scope.
AGEI Support Today:
Risk classification enum, HITL risk triggers, tool risk classes, gate evaluation evidence.
Current Status:
Strong request/tool-level support; system-level risk register still needed.
ADG Expects: No self-approval; clear authority boundaries across Adopt, Defend, Govern.
AGEI Support Today:
Server-side gates, approve/deny/escalate/inspect outcomes, HITL decisions, reviewer grants, authorization audit events, agent delegations.
Current Status:
Strong current support through server-side gates, HITL evidence, reviewer grants, and authorization audit events.
ADG Expects: Evaluation before deployment, including safety, security, quality, fairness, and approval evidence.
AGEI Support Today:
Gate definitions, gate evaluations, policy versions, artifact release records, pre-action proof bundles.
Current Status:
Strong foundation — ADG-style release checklist template needed.
ADG Expects: Governed changes to models, prompts, tools, policies, retrieval sources, and configurations.
AGEI Support Today:
Policy versions, lifecycle objects, artifact release records, receipts.
Current Status:
Versioning exists; consolidated change log artifact needed.
ADG Expects: Context provenance, retention, access restrictions, trust ordering, and privacy controls.
AGEI Support Today:
Evidence capsules, privacy treatments, retention policies, personal data references, consent records, redaction events.
Current Status:
Schema-supported — context policy artifact and UI needed.
ADG Expects: Tool and MCP registry with trust tiers, invocation controls, and audit logging.
AGEI Support Today:
Agent tool definitions, tool invocations, required gate definitions, shadow AI tool registry, service grants.
Current Status:
Strong schema support — Tool/MCP Register UI and export needed.
ADG Expects: Runtime monitoring for abuse, drift, unsafe actions, leakage, bias emergence, and configuration drift.
AGEI Support Today:
Gate evaluations, receipts, agent sessions, tool invocations, behavior baselines, anomaly alerts, API request logs.
Current Status:
Strong data capture — monitoring dashboard needed.
ADG Expects: AI-specific incident response with replayable evidence capture.
AGEI Support Today:
Incidents, incident evidence links, audit packs, materialization requests, receipts.
Current Status:
Schema exists; incident workflow UI and incident evidence pack template needed.
ADG Expects: Periodic review records, exception handling, and board reporting.
AGEI Support Today:
Audit packs, verification jobs, agent access reviews, authorization events, policy history.
Current Status:
Governance review cycle and exception register needed.
ADG Expects: Fairness evaluation reports and re-evaluation after material changes.
AGEI Support Today:
DPIA records, gate evaluation evidence, policy templates can store bias metrics.
Current Status:
Schema-supported — dedicated fairness workflow and report template needed.
ADG Expects: Vendor accountability boundaries, assurance records, and shared responsibility matrix.
AGEI Support Today:
Processing activities, organization members, reviewer grants, agent delegations, authorization events.
Current Status:
Schema-supported — vendor due diligence and shared responsibility artifact needed.
Every governed event creates cryptographic receipt
Approve, deny, escalate, inspect decisions
Human-in-the-loop decision tracking
Evidence chain connections
Evidence bundle creation and sealing
Independent receipt verification
Permission and access tracking
Tool registry and invocation logging
Unapproved AI/tool usage reporting and evidence capture
GDPR compliance tracking
Current roadmap items to strengthen ADG alignment
AGEI support for the 11 ADG required operating artifacts
| Artifact | AGEI Status | Details |
|---|---|---|
| AI System Profile | Partial | Data exists, export needed |
| ADG RACI Matrix | Roadmap | Template needed |
| Agent Authority Statement | Partial | Delegation/boundary data exists, export needed |
| Context Policy | Roadmap | Schema-supported |
| Tool and MCP Register | Partial | Schema exists, UI/export needed |
| AI Release Gate Checklist | Partial | Gate data exists, template needed |
| AI Incident Evidence Pack | Partial | Incident/audit pack data exists, template needed |
| Model and Prompt Change Log | Partial | Policy versioning exists, prompt/model change log needed |
| Governance Exception Register | Roadmap | Not yet implemented |
| Vendor AI Due Diligence Record | Roadmap | Template needed |
| Fairness Evaluation Report | Roadmap | Template needed |
ADG defines HITL, HOTL, and HOOTL as operational autonomy tiers
Human approves every action before execution
Supported through HITL requests, HITL decisions, escalate gate outcome, and human review evidence capture.
Human monitors and can intervene, but action proceeds
Partially supported through webhooks, anomaly alerts, and monitoring evidence. Explicit "monitoring mode" configuration is a roadmap item.
Fully autonomous with retrospective review only
Partially supported through approve gate outcomes and post-decision audit packs. Dedicated retrospective review workflow is a roadmap item.
AGEI provides governance evidence infrastructure. It does not provide legal advice, audit certification, ADG certification, or a guarantee of regulatory compliance. Evidence generated through AGEI may support governance, audit, risk, legal, and compliance review when the system is configured and used correctly.