1. Background

1.1 Origin and Positioning

ADG began with a mission at EC-Council: the AI governance space was fragmenting into point standards, vendor playbooks, and academic frameworks - none of which gave enterprise boards and engineering teams a single, practitioner-tested operating model for deploying AI safely at scale. EC-Council set out to build that missing standard: a framework built around three enduring functions - ADOPT DEFEND GOVERN - that scales from the board to the engineer.

To avoid building in a vacuum, EC-Council convened an Advisory Board of practitioners from mature enterprises that have actually deployed AI in production. The initial draft was circulated to senior AI, security, and governance leaders across global organizations in financial services, technology, manufacturing, healthcare, telecommunications, energy, and consulting - covering both regulated and less-regulated sectors, and spanning North America, Europe, and Asia-Pacific.

Framework Leadership

Jay Bavisi

Chairman & CEO, EC-Council Group

Karthik S.

Framework Architect & Lead Author

Practice Head, SecureAI, EC-Council Global Services

Mayank Tandon

Global Outreach & Partner Experience

EC-Council

With Gratitude

Thank you to the Advisory Board

The practitioners below gave their time, scrutiny, and hard-won experience to review ADG. Their comments, corrections, and counterpoints are the reason this framework is field-tested rather than aspirational.

Adam Spearing

VP of AI GTM EMEA

ServiceNow

Andrei Son

Head of AI Transformation

AUMOVIO

Anish Mitra

Director

KPMG

Anita Lacea

Head of AI Transformation, Azure Hardware Infrastructure

Microsoft

Dinesh Bhogle

Head of AI/ML platform

Black & Veatch

Dr. Sayed Peerzade

Executive Vice President — Cloud, AI & Government Initiatives

Jio

Edoardo Tealdi

Executive Head of AI Transformation — Business Engagement and Growth Units

NTT DATA, Inc.

Kathy Baxter

VP / Principal Architect, Responsible AI & Tech

Salesforce

Lewis V. Adams

Vice President | Enterprise AI & Capital Productivity Transformation

Citi

Lily Rachmawati

Director, Head of Applied AI

BNP Paribas

Malik Hussain

AI Enablement Lead, Data & AI Academy

BASF

Mark Ritcey

Vice President, AI & Automation Delivery

Latent Bridge

Naveen Upadhyay

Vice President, AI/ML Product Management – Machine Learning & Intelligence Operations

JPMorgan Chase & Co.

Oscar Jarabo

Global Head of AI Product & Strategy

TKE

Pavan Kristipati

Head of AI Engineering & Transformation (SVP Scope) | Enterprise AI Adoption, Governance & Platform

Huntington Bank

Raghunandan Mishra

AI & Data Engineering Leader

Raji Bhimireddy

Vice President Cloud, AI, Architecture, FInOps & Business Value

Prudential

Sanjoy K. Saha

Head of AI Portfolio and Governance & Chief of Staff CDAO

GE Healthcare

ShanShan Pa

Global Head of AI & Data Governance

GlobalLogic

Sophia Katrenko

VP of AI/ML

EcoVadis

Sruthi Pakanati

Head of AI & Data Transformation, National Quality & Risk

Deloitte Australia

Sudarson Roy Pratihar

Founder and Principal

A2IQ

Yashwinder Chhikara

Sr. Vice President — AI, Analytics, and Product Management

iSON Xperiences

This version is the result - the iterative output of that collaboration. ADG is not a one-and-done document. It is a living framework, designed to be enhanced as AI deployment patterns evolve and as new practitioners join the conversation. The scope has grown from its original LLM-centric framing to cover modern deployment patterns including LLMOps, retrieval-augmented systems, tool-using agents, MCP-connected ecosystems, long-context architectures, multi-agent systems, and model lifecycle controls spanning pre-training through runtime.

Practitioners who want to contribute to the next version can join the Advisory Board here.

1.2 Framework Purpose: The RE³ Trust Model

ADG is purpose-built to operationalize four non-negotiable properties of trustworthy enterprise AI - captured in the RE³ Trust Model:

These are not checkboxes. They are governance outcomes that must be engineered into every AI deployment - from an internal copilot through a production multi-agent system. ADG translates each outcome into concrete controls across all nine governance surfaces (Model, Prompt, Context, Tools, Orchestration, Identity, Safety Layer, Telemetry, Learning Loop) and applies them to every deployment pattern (LLMOps, Agentic, Tools & MCP, Context, Training, Multi-Agent, Multi-Modal, and composite systems) across all three deployment classes (Homegrown, Foundation Model API, SaaS AI).

The outcome: RE³ AI becomes measurable. Each trust property maps to specific controls, metrics, and evidence requirements, so organizations can demonstrate trustworthy AI to boards, regulators, and customers - not just claim it.

1.3 Core Thesis

Traditional cyber and software governance frameworks assume deterministic execution, finite input spaces, stable functionality, and visible logic. AI systems violate those assumptions. Organizations need a framework that governs:

• Systems that can generate novel behavior
• Systems that act through external tools and APIs
• Systems that consume dynamic context from internal and external sources
• Systems whose behavior can change through model updates, tuning, retrieval changes, and runtime interaction
• Systems composed of multiple models and agents that interact with each other, creating emergent risks not present in any individual component

ADG addresses this by combining role separation, surface-based control ownership, deployment-pattern overlays, lifecycle governance, and a structured harm taxonomy that extends beyond security to encompass responsible AI.

1.4 Scope and System Coverage

ADG governs all AI systems as defined by the OECD AI Policy Observatory and the EU AI Act: machine-based systems that, for explicit or implicit objectives, infer from input how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. This explicitly includes:

• Large language models (LLMs) and transformer-based generative AI
• Diffusion models, diffusion-transformer hybrids, and image/video generation systems
• Multi-modal systems combining text, image, audio, and code generation
• Classical machine learning models (supervised, unsupervised, reinforcement learning)
• Composite AI systems that orchestrate multiple model types within a single workflow
• Emerging architectures including world models, neuromorphic systems, and foundation models beyond text

The governance surfaces, overlays, and controls in this framework are architecture-neutral by design. Where a deployment pattern creates unique governance requirements (e.g., agentic orchestration, diffusion-model provenance), a dedicated overlay provides additional controls. The framework is not limited to LLM-centric systems.

1.5 Distinction: Standards vs. Regulations

ADG is a voluntary governance framework designed to be compatible with, but distinct from, regulatory requirements. Throughout this document:

• Regulatory requirements refer to legally binding obligations (EU AI Act, sector-specific regulations, data protection laws such as GDPR)
• Standards refer to voluntary frameworks and best practices (NIST AI RMF, ISO/IEC 42001, OWASP Top 10 for LLM Applications, OWASP Top 10 for Agentic Applications, MITRE ATT&CK, MITRE ATLAS)
• Threat catalogues describe what attackers do. ADG consumes four canonical sources - MITRE ATT&CK, MITRE ATLAS, OWASP Top 10 for LLM, OWASP Top 10 for Agentic - and maps each technique to the Minimum Control Set and on to NIST / ISO. See the §11.2.1 interactive flow.

ADG controls are designed to satisfy both regulatory and standards-based requirements, but organizations must independently verify compliance with applicable legal obligations in their jurisdictions.

2. Executive Summary

ADG is an enterprise AI security and responsible AI governance framework structured around three pillars - ADOPT DEFEND GOVERN - that provide a board-to-engineering operating model for safe, responsible, and scalable AI deployment.

Key framework components:

This version was forged with senior AI, security, and governance leaders running production AI inside Fortune 500, Fortune Global 500, and Big Four firms - practitioners working from design through implementation across regulated and less-regulated sectors. It addresses eight major gap clusters identified through systematic review and extends the framework to cover multi-agent systems, multi-modal architectures, shared responsibility, responsible AI, and measurable governance. It is suitable as the basis for consulting assessments, enterprise AI governance programs, certification architecture, board-facing assurance discussions, regulatory preparation, and vendor due diligence.

3. Document Usage

This section provides guidance on how to read, apply, and implement the ADG framework.

3.1 Target Audience

Each audience has a tailored reading path. Click any section to jump there directly.

3.2 Foundational Principles

ADG is built on nine foundational principles. Each principle is a durable standard that applies across industries, deployment patterns, and regulatory regimes.

3.3 Autonomy Tiers

ADG distinguishes three operational autonomy tiers. The tier determines the minimum governance requirements for deployment.

Note on terminology: HITL/HOTL/HOOTL originated in DoD doctrine (DoD 5000.59-M, 1998) and remain widely used, though their boundaries are debated in current AI-safety literature. ADG uses the tighter operational definitions below to make each tier directly testable against the Minimum Control Set.

3.4 Implementation Roadmap Overview

4. The ADG Triad and Operating Model

4.4 Operating Model

ADG uses a simple rule: ADOPT builds and operates, DEFEND breaks and protects, GOVERN authorizes and oversees.

5. Governance Surfaces

ADG organizes governance into nine surfaces. These surfaces define what must be governed regardless of model vendor, model architecture, or deployment pattern.

6. Harm Taxonomy and Responsible AI Integration

ADG recognizes that AI governance must address harms beyond security vulnerabilities. The framework adopts a four-class harm taxonomy that maps into every governance surface, lifecycle stage, and deployment overlay.

6.1 Harm Classification

The four harm classes plotted by detection difficulty and impact scope. Each quadrant carries the threats it covers, the pillars that own it, and the Minimum Controls that detect it.

6.2 Responsible AI Principles

Every governance surface in ADG must account for seven responsible AI dimensions, embedded across all three pillars rather than appended as a separate track.

Integration rule: Responsible AI is not a separate track. It is embedded into every ADG pillar - ADOPT, DEFEND, and GOVERN - and must be reflected in all certification curricula, operating artifacts, and governance reviews.

7. Shared Responsibility Model

When organizations deploy AI, responsibility for governance controls varies based on how the AI is provisioned. ADG defines three deployment responsibility classes and maps control ownership for each.

7.1 Deployment Responsibility Classes

7.2 Vendor AI Due Diligence Requirements

For Foundation Model API and SaaS AI deployments, organizations must:

1. Require vendor disclosure of model provenance, training data policies, alignment methods, and known limitations
2. Establish contractual AI governance clauses covering incident notification, data handling, model change management, and liability allocation
3. Conduct independent evaluation of vendor-provided safety controls and not rely solely on vendor claims
4. Maintain consumer-side telemetry and monitoring regardless of vendor monitoring capabilities
5. Include AI governance questions in procurement processes and vendor risk assessments
6. Define rollback and exit strategies for vendor AI dependencies

8. People, Process, Technology, and Data

8.1 People Layer

ADG requires explicit capability ownership in addition to functional ownership. The people model includes dedicated responsible AI expertise across all pillars.

• ADOPT roles: AI product owner, accountable AI service owner, LLMOps lead, orchestration engineer, prompt designer, platform engineer, enterprise architect, DevSecOps engineer
• DEFEND roles: AI red team lead, guardrail engineer, detection engineer, AI security architect, incident commander, forensic analyst, responsible AI scientist, bias/fairness auditor
• GOVERN roles: model risk officer, privacy counsel, compliance lead, data steward, AI ethics and sociotechnical oversight lead, executive approver, procurement governance lead

AI Governance Council: Every organization deploying AI systems should establish a cross-functional AI Governance Council (or equivalent steering body) with representation from ADOPT, DEFEND, and GOVERN. This council serves as the escalation path and tension-resolution mechanism described in Section 4.4.

8.2 Process Layer

ADG formalizes the minimum process backbone required for enterprise implementation:

1. AI system inventory and criticality classification
2. Deployment gate and approval workflow (including pre-deployment fairness and safety evaluation)
3. Prompt, model, and tool change management
4. Continuous evaluation, drift review, and fairness monitoring
5. Human oversight model selection: HITL, HOTL, or HOOTL with documented rationale
6. AI-specific incident response and evidence retention
7. Post-incident remediation with retest and governance review
8. Vendor AI due diligence and ongoing assurance
9. Board-level reporting cycle for high-risk AI systems

8.3 Technology Layer

ADG treats the following as governed technology layers:

1. Model gateway and routing layer
2. Retrieval, memory, and context management layer
3. Tool registry and MCP trust layer
4. Policy engine, guardrail layer, and harm detection layer
5. Runtime telemetry, replay, and fairness measurement layer
6. Training, tuning, and evaluation pipeline layer
7. Input validation and sanitization layer (covering prompts, RAG inputs, and multi-modal inputs)

8.4 Data Layer

Traditional enterprise data architecture separates data into distinct tiers - structured databases, data warehouses, data lakes, unstructured file stores, and APIs - each with its own governance model, access controls, and tooling. AI systems fundamentally disrupt this separation. When an AI agent consumes enterprise data through retrieval pipelines, MCP connections, or tool invocations, it does not distinguish between a SQL database record, a SharePoint document, a Slack thread, or a PDF. All data is converted into text tokens within a context window, collapsing traditional tier boundaries into a single consumption surface.

This convergence means that AI systems can reassemble sensitive information from fragments scattered across sources that were never intended to be combined, infer personal data from context that contains no explicit PII, and traverse entire enterprise file systems where content is readable as text. The governance implication is profound: access control at the storage layer alone is no longer sufficient. Organizations must govern the full pipeline from source data through retrieval, embedding, context assembly, and AI consumption.

ADG requires organizations to govern data across four dimensions:

8.4.1 Data People

• Data Stewards for AI: extending traditional data stewardship to govern the full text-based data surface that AI systems can access
• AI Data Engineers: bridging data engineering and AI operations, responsible for RAG indexing, embedding generation, knowledge base curation
• Context Architects: designing what data flows into AI context windows, in what priority order, with what trust ranking
• Knowledge Base Curators: responsible for freshness, accuracy, deduplication, and retirement of enterprise knowledge assets

8.4.2 Data Process

• AI-aware data classification: extending beyond storage-tier access control to include AI-readability rules
• Provenance tracking across the text pipeline: maintaining a verifiable chain from any AI output back through the retrieval step to the source document
• Cross-source inference governance: policies governing when AI systems may combine information from multiple data sources
• Text-based policy management: AI system configurations governed as controlled documents with versioning, approval workflows, and rollback
• Knowledge base lifecycle management: content ingestion, quality validation, freshness review, conflict resolution, deduplication, and retirement
• Data minimization for AI: ensuring context windows contain only data necessary for the task
• Consent and lawful basis tracking: maintaining records of lawful basis for processing each data category

8.4.3 Data Technology

• Vector stores and embedding infrastructure: governed data infrastructure requiring access controls, encryption, backup and recovery
• Enterprise knowledge graphs: structured representations enabling context assembly with awareness of entity relationships
• Context assembly engines: systems that select, rank, truncate, filter, and compose text from multiple sources
• MCP and tool registries as data access layers: governed as data access infrastructure with the same rigor as database connections
• Text-based configuration stores: GitOps-style repositories for all AI system configurations
• Data lineage and output attribution: technology to trace which source documents contributed to a specific AI output
• Embedding pipeline governance: validation testing, drift monitoring, and re-indexing governance

8.4.4 Data Governance Integration with ADG Surfaces

9. Lifecycle Governance

ADG applies controls across six lifecycle stages. The lifecycle is explicitly circular: lessons from Retire/Learn feed back into Pre-training/Sourcing decisions.

9.1 Post-Deployment Continuous Governance

The Run/Monitor stage requires specific continuous governance controls that go beyond traditional operational monitoring:

1. Configuration drift detection: verify that the system in production matches what was approved (models, prompts, tools, context sources)
2. Usage authorization monitoring: confirm that approved users are using the system for approved purposes within approved boundaries
3. Feature and capability change governance: new model versions, prompt updates, tool additions, and retrieval source changes post-deployment must go through change control
4. Continuous automated evaluation: scheduled adversarial testing, fairness benchmarking, and accuracy regression testing on production systems
5. Ongoing governance of product roadmap: feature additions to deployed AI systems require re-evaluation against the original risk classification and approval conditions

10. Deployment Pattern Overlays

ADG adds deployment-specific overlays so that governance matches the actual architecture being deployed. Overlays are additive - select all that apply.

10.1 LLMOps Overlay

Scope: versioning, evaluation, release management, rollback, cost control, and performance monitoring for LLM-based services.

• ADOPT structures model onboarding, baseline evaluation, config versioning, rollback readiness, and operational SLOs
• DEFEND structures adversarial testing, abuse simulation, leakage testing, denial-of-wallet controls, monitoring coverage, and bias evaluation
• GOVERN structures use-case approval, vendor due diligence, deployment thresholds, exception management, and LLMOps cost governance

10.2 Agentic Orchestration Overlay

Scope: planners, loops, retries, multi-step reasoning, subtask decomposition, and multi-agent coordination.

• ADOPT defines mission scope, stop conditions, retry budgets, and workflow boundaries
• DEFEND validates loop abuse resistance, prompt chaining resistance, recursion limits, and kill-switch behavior
• GOVERN approves autonomy tier, escalation path, and legal accountability for delegated decisions

10.3 Agent Overlay

Scope: AI systems that act on behalf of a user, team, or enterprise process rather than merely generating content.

• ADOPT defines business mission, action boundaries, and acceptable failure modes
• DEFEND tests transaction safety, impersonation resistance, and unsafe action prevention
• GOVERN defines liability model, mandatory approvals, record retention, and segregation-of-duty requirements

10.4 Agentic Hardening Overlay

Required control themes:

1. Delegated authority bounds
2. Action-layer isolation and dry-run modes
3. Semantic firewalls before tool use and before response release
4. Memory hygiene and state reset controls
5. Human override and emergency stop pathways
6. Forensic replay of prompts, retrieved context, policy decisions, and tool calls

10.5 Tools and MCP Overlay

Scope: tool discovery, capability registration, trust mediation, and policy-controlled invocation of tools and context providers.

• ADOPT justifies why each tool or MCP capability is needed and what business task boundary it serves
• DEFEND enforces authentication, authorization, parameter validation, rate limiting, sandboxing, and audit logging
• GOVERN maintains approval policy, third-party assurance criteria, trust tiers, and data-sharing restrictions

Minimum MCP requirements:

1. Approved server inventory
2. Capability trust tiering
3. Per-invocation policy evaluation
4. Auditable request and response traces
5. Explicit rejection of implicit trust in tool-supplied instructions

10.6 Context and Long-Window Overlay

Scope: session history, retrieved enterprise knowledge, hidden orchestration instructions, persistent memory, and context overflow handling.

• ADOPT defines the minimum context required for task quality
• DEFEND tests for context poisoning, leakage, overflow abuse, cross-session contamination, and multi-turn attack patterns
• GOVERN defines provenance rules, retention, data-class handling, trust ranking, and lawful-use constraints

Minimum context policy: every production system must define approved sources, retention periods, redaction rules, persistence rules, and trust ordering across context inputs.

10.7 Pre-training and Post-training Overlay

Scope: model provenance, fine-tuning, adapters, preference alignment, safety tuning, feedback loops, and retraining.

• ADOPT structures fine-tuning objectives, experiment tracking, rollback, and operational performance baselines
• DEFEND structures regression testing, jailbreak retesting, alignment failure analysis, fairness drift monitoring, and bias evaluation
• GOVERN structures provenance review, licensing, cross-border constraints, GDPR compliance for training data, documentation, and approval of learning-loop inputs

10.8 Multi-Agent and Agent Interoperability Overlay

Scope: systems where multiple AI agents interact, coordinate, negotiate, or delegate tasks to each other.

Multi-agent systems create emergent governance challenges that single-agent controls do not address:

• Agent-to-agent trust boundaries: each agent interaction must enforce explicit trust verification; no agent should implicitly trust another agent's outputs or instructions
• Value alignment verification: agents with different guardrail configurations, different training, or different vendors may have incompatible safety policies. Cross-agent interactions must be tested for value alignment conflicts
• Collusion and deception detection: monitoring for patterns where agents coordinate to bypass controls, produce misleading outputs, or exploit gaps between their respective guardrails
• Cascading failure and amplification risk: a failure or bias in one agent can be amplified through a chain of dependent agents. Circuit breakers must exist at agent-to-agent boundaries
• Cross-agent audit trails: every agent-to-agent interaction must be logged with sufficient detail for forensic replay
• Sycophancy degradation prevention: in agent-agent interactions, sycophantic behavior can devolve into harmful feedback loops. Detection and interruption mechanisms are required
• Interoperability governance: when agents from different organizations or vendors interact, a shared governance protocol must define minimum safety, logging, and accountability requirements

10.9 Multi-Modal and Composite AI Systems Overlay

Scope: systems combining multiple model architectures (LLM + diffusion + classifier + retrieval) within a single product or workflow.

• ADOPT defines the system-level architecture, ensuring each model component has a clear purpose and integration specification
• DEFEND tests for cross-model interaction risks: outputs from one model becoming adversarial inputs to another, amplification of individual model weaknesses through the chain
• GOVERN requires system-level risk assessment, not just per-model evaluation, and defines composite system approval criteria

11. Controls, Artifacts, and Measurement

11.1 Minimum Control Set

The following controls form the minimum baseline for ADG implementation. Each control includes an evidence requirement to ensure measurability.

11.2 Control Crosswalk - NIST AI RMF & ISO/IEC 42001

Each Minimum Control maps to one or more NIST AI RMF functions and ISO/IEC 42001 Annex A groups. Click any control below to reveal how many domains it touches across both frameworks.

11.2.1 Threat → Control → Standard Flow

The crosswalk above shows ADG controls mapping to NIST and ISO. The flow below adds the technical layer on top: every threat catalogued by MITRE ATT&CK, MITRE ATLAS, OWASP Top 10 for LLM (2025), and OWASP Top 10 for Agentic (2026) flows into one or more ADG Minimum Controls - and from there to the NIST AI RMF function and ISO/IEC 42001 Annex A clause that recognises the same control objective. Pick a catalogue tab, click a threat, and trace the chain.

11.3 Required Operating Artifacts

11.4 Measurement and Evidence Framework

ADG requires measurable governance. Three measurement tiers ensure controls are not just documented but demonstrably effective.

12. Regulatory Alignment and Roadmap

12.1 Regulatory and Standards Alignment

ADG is designed to be compatible with major AI governance regulations and standards. The following matrix maps ADG components to key external frameworks.

Note: This mapping is indicative, not exhaustive. Organizations must conduct their own regulatory compliance assessment. ADG provides a governance backbone that facilitates compliance but does not guarantee it.

For the full 12-control Minimum Control Set mapped to NIST AI RMF and ISO/IEC 42001, see §11.2 Control Crosswalk. The table below covers EU AI Act alignment and cross-cutting ADG components.

12.2 Implementation Roadmap

13. Workforce Capability

13.1 The AI Skills Gap

The defining constraint on enterprise AI is not the technology - it is the workforce. AI is scaling into production faster than the people who must run, secure, and govern it can be trained, credentialed, and deployed. IDC projects that by 2026 more than 90% of organizations will feel the IT skills crisis, at a cost of $5.5 trillion in delays, quality issues, and lost competitiveness (IDC, Enterprise Resilience: IT Skilling Strategies, 2024). Without workforce capability, every control in this framework remains theoretical.

"AI is moving from experimentation to infrastructure, and the workforce has to move with it. Security leaders are now accountable for systems that learn, adapt, and influence outcomes at speed."
- Jay Bavisi, Chairman & CEO, EC-Council Group

The ADG framework defines three pillars - ADOPT DEFEND GOVERN - and each one surfaces a distinct capability gap that is showing up across enterprises today:

• The Adopt gap: program and transformation leaders cannot translate AI strategy into measurable, governed delivery - stalling pilots before they reach production.
• The Defend gap: security organisations lack adversarial AI expertise - prompt injection, model evasion, data poisoning, and multi-agent exploitation are not yet standard tradecraft.
• The Govern gap: risk, compliance, and board leaders lack the vocabulary, evidence protocols, and regulatory mapping (EU AI Act, NIST AI RMF, ISO/IEC 42001) to hold AI systems to account.

13.2 The Capability Bridge

EC-Council's answer is the Enterprise AI Credential Suite - a portfolio of role-aligned credentials mapped directly to the ADG framework. Launched February 2026 as the largest single portfolio expansion in EC-Council's 25-year history, the suite is structured as a four-part capability ladder:

1. Literacy baseline - a shared AI vocabulary for every practitioner, technical or not, so governance conversations can happen at enterprise speed. Anchored by the Artificial Intelligence Essentials (AIE) programme.
2. Role-aligned capability - three flagship credentials mapped 1:1 to the ADG pillars: C|AIPM for Adopt, C|OASP for Defend, C|RAGE for Govern. Each credential certifies execution-grade competence for the people accountable for that pillar.
3. Extended core skills - CEH v13 restructured to integrate AI modules across all five phases of ethical hacking, extending existing security practitioners into the AI attack surface without a role change.
4. Operational readiness - each credential crosswalks to the framework's Minimum Control Set (MC-1 through MC-12), turning certification into deployable evidence that auditors, regulators, and boards can accept.

13.3 Credential Portfolio

The five credentials below map to the four-part capability ladder in Section 13.2. Each links to its official EC-Council programme page for syllabus, prerequisites, and enrolment.

The full EC-Council AI credential catalogue is maintained at eccouncil.org/ai-courses.

Appendix A: Definitions

Executive Conclusion

ADG turns the original ADG concept into a full enterprise AI security and responsible AI governance framework. It keeps the simplicity of ADOPT, DEFEND, and GOVERN, but adds the structure required to govern contemporary AI systems - across people, process, technology, deployment pattern, lifecycle, harm class, and regulatory environment.

This version was forged with senior AI, security, and governance leaders running production AI inside Fortune 500, Fortune Global 500, and Big Four firms - practitioners working from design through implementation across regulated and less-regulated sectors. It addresses eight major gap clusters identified through systematic review, and extends the framework to cover multi-agent systems, multi-modal architectures, shared responsibility, responsible AI, and measurable governance.

From a framework standpoint, this version is suitable as the basis for:

• Consulting assessments and maturity reviews
• Enterprise AI governance programs
• Certification architecture and training design across all three ADG pillars
• Board-facing AI security, responsible AI, and assurance discussions
• Regulatory preparation and compliance gap analysis
• Vendor AI due diligence and procurement governance

ADG | Adopt | Defend | Govern - AI Security Governance Framework

Enhanced Working Draft • April 2026 • Confidential