Your compliance team has been working on EU AI Act readiness for eighteen months. Your security team adopted NIST AI RMF as an internal risk vocabulary eighteen months before that. Last year, someone in procurement decided ISO/IEC 42001 certification was becoming a customer requirement. Now you have three parallel workstreams, three documentation formats, three sets of internal review meetings — and the same underlying AI inventory, risk assessments, and governance policies sitting in different places across all three programs.
This is not a compliance problem. It is an architecture problem. And it is the most common structural failure in enterprise AI governance today.
This article is a blueprint for building one unified AI governance operating model that satisfies the EU AI Act, maps to the NIST AI RMF, and provides the management system infrastructure required for ISO/IEC 42001 certification — simultaneously, from a single set of documented controls. It covers how the three frameworks differ, where they converge, what genuinely requires separate treatment, and how to sequence implementation for maximum efficiency.
Why Framework Convergence Is No Longer Optional
AI governance has moved from voluntary best practice to binding legal obligation in less than two years. The EU AI Act entered into force on August 1, 2024. Prohibited AI practices became enforceable on February 2, 2025. General-purpose AI model obligations applied from August 2, 2025. High-risk AI system obligations — the most operationally demanding wave of the regulation — are enforceable from August 2, 2026, with penalties reaching up to €35 million or 7% of global annual turnover for the most serious violations.
The regulation is explicitly extraterritorial. Any organization placing AI systems on the EU market, or deploying AI whose outputs affect EU residents, must comply regardless of where it is headquartered. That scope brings in a substantial share of the global enterprise AI estate. Many of those same organizations are simultaneously responding to procurement questionnaires, customer due diligence requests, and insurance underwriting requirements that reference NIST AI RMF alignment or ISO/IEC 42001 certification. The regulatory mandate and the market mandate are arriving together.
The fragmentation risk is real and measurable. Organizations that run separate compliance tracks for each framework duplicate risk assessment processes, maintain inconsistent documentation of the same underlying AI systems, and create audit fatigue that degrades quality over time. When a high-risk AI system requires an Article 9 risk management record under the EU AI Act, an ISO 42001 Clause 6.1 risk assessment, and an AI RMF Measure function output — and these three documents say different things because they were prepared independently — the organization has not only wasted resources, it has created an evidentiary problem if regulators or auditors ever compare them. The operational reality of running enterprise AI compliance in 2026 is that governance fragmentation is itself a risk, not a precaution.
The macro pressure is also accelerating. Boards are asking for AI governance accountability. Investors are treating AI risk management as a component of ESG and enterprise risk disclosure. Procurement teams at large enterprises are issuing AI governance questionnaires to suppliers as standard commercial practice. The question is not whether your organization needs a governed AI program. The question is whether you can sustain three separate governance programs as your AI estate grows — or whether a unified architecture is the only scalable path.
Understanding the Three Frameworks
The EU AI Act, the NIST AI RMF, and ISO/IEC 42001 occupy different positions in the governance landscape and were designed to do different things. Understanding those differences precisely is the foundation for combining them intelligently.
The EU AI Act: Binding Regulatory Obligation
The EU AI Act (Regulation 2024/1689) is the world's first comprehensive horizontal AI law. It takes a risk-tiered approach: AI systems posing unacceptable risk are prohibited outright; high-risk systems face extensive mandatory obligations; limited-risk systems face transparency duties; and minimal-risk systems are unregulated. The eight categories of high-risk AI under Annex III include employment screening, credit scoring, biometric identification, educational assessment, critical infrastructure management, law enforcement tools, migration and border control, and systems affecting access to essential services. If your AI system falls into any of these categories and its outputs touch the EU, you have binding legal obligations with enforcement teeth.
For providers — organizations that develop high-risk AI systems — the obligations are comprehensive: a documented risk management system under Article 9, training data governance under Article 10, technical documentation under Article 11 and Annex IV, automatic logging under Article 12, transparency obligations under Articles 13 and 14, human oversight mechanisms under Article 14, accuracy and cybersecurity controls under Article 15, a quality management system under Article 17, conformity assessment before market entry, CE marking, and registration in the EU AI database under Article 49. For deployers — organizations that use but did not develop a high-risk system — the obligations are less extensive but still operational: using systems according to instructions, maintaining human oversight, retaining logs for at least six months, and conducting Fundamental Rights Impact Assessments where required.
The penalties are what make this framework categorically different from the others. Up to €35 million or 7% of global annual turnover for violations of prohibited AI practice prohibitions. Up to €15 million or 3% for non-compliance with high-risk system obligations. National market surveillance authorities can also withdraw non-compliant AI systems from the EU market, which for organizations whose core operations depend on that AI represents a commercial risk that dwarfs the financial penalty.
NIST AI RMF: Voluntary Risk Management Framework
The NIST AI Risk Management Framework, published by the US National Institute of Standards and Technology in January 2023 and expanded through 2024–2025 with the Generative AI Profile (AI 600-1) and additional companion playbooks, is a voluntary, sector-agnostic framework for managing AI risk. It does not carry legal force, does not require certification, and imposes no penalty for non-adoption. What it provides is a structured, operationally detailed governance vocabulary organized around four core functions: Govern, Map, Measure, and Manage.
Govern establishes organizational accountability, policies, and culture for AI risk. Map characterizes AI systems, their contexts, and the risks they pose. Measure analyzes and assesses those risks using quantitative and qualitative methods. Manage develops response plans, applies controls, and monitors residual risk. These functions are intended as an iterative cycle, not a linear process — AI systems change after deployment, and governance must track those changes. The framework's strength is in its operational granularity. The accompanying AI RMF Playbook maps each function to specific suggested actions, and NIST has published a formal crosswalk between the AI RMF and ISO/IEC 42001 that explicitly maps AI RMF subcategories to ISO 42001 clauses. That crosswalk is the starting point for any unified governance architecture.
The AI RMF is increasingly referenced by US regulators, federal procurement requirements, and sector-specific guidance (healthcare HIPAA, financial services model risk management) as the expected baseline for responsible AI governance. Even as a voluntary framework, non-alignment creates reputational and procurement risk in regulated sectors.
ISO/IEC 42001: Certifiable AI Management System Standard
ISO/IEC 42001, published by the International Organization for Standardization in December 2023, is the world's first international standard for an AI Management System. Unlike the AI RMF, it is designed for third-party certification. Unlike the EU AI Act, it is not a regulation. Its purpose is to provide a structured, certifiable evidence base that an organization's AI governance is systematic, documented, and continuously improving — the same model that made ISO 27001 for information security and ISO 9001 for quality management global governance benchmarks.
The standard uses the high-level structure common to all ISO management system standards, which means organizations already certified to ISO 27001 can integrate AI governance into an existing management system infrastructure with substantially reduced implementation effort. The audit cadence — Stage 1 documentation review, Stage 2 on-site assessment, annual surveillance audits, full recertification in year three — is familiar to any organization with existing ISO certifications. The Annex A controls address AI impact assessment, data governance, AI system lifecycle management, transparency, human oversight, and stakeholder engagement. Importantly, ISO 42001's Clause 4.1 explicitly cross-references the NIST AI RMF for AI role types and lifecycle stages, signaling that the two frameworks were designed to be read together.
The commercial case for ISO 42001 certification is growing rapidly. Enterprise procurement teams, insurers, and regulators are treating certification as a trust signal for AI governance maturity in a way that self-attested NIST AI RMF alignment cannot replicate. ISO/IEC 42001 certification is also emerging as documentary evidence for regulators evaluating compliance with the EU AI Act's Article 17 quality management system requirement — not a formal substitute for conformity assessment, but a strong signal of systematic governance.
For a detailed view of how AI governance applies at each stage of the AI lifecycle — and how these framework functions map onto concrete engineering and compliance activities — the lifecycle structure is the clearest organizing principle.
Where the Three Frameworks Converge
The most operationally valuable insight for building a unified governance program is understanding what these frameworks actually share — because the overlap is substantial, and shared requirements are the building blocks of a single evidence base.
All three frameworks require an AI inventory as a foundational governance artifact. You cannot manage risk you cannot see. The EU AI Act requires providers to maintain technical documentation and deployers to understand what systems they are using. The AI RMF's Map function begins with characterizing AI systems and their contexts. ISO 42001 requires organizations to define the scope of the AI management system and understand the AI systems within it. The AI inventory — a register of every AI system in operation, with information on purpose, data inputs, decision outputs, affected populations, risk classification, and deployment context — is a shared foundational document across all three frameworks. Built once, maintained continuously, it serves all three.
Risk assessment is similarly shared. The EU AI Act's Article 9 risk management system, ISO 42001's Clause 6.1 risk assessment, and the AI RMF's Measure function all require organizations to identify AI risks, evaluate their likelihood and impact, and document mitigation controls. The risk categories they focus on differ in emphasis but overlap significantly: accuracy and reliability failures, bias and discrimination, transparency and explainability gaps, security vulnerabilities, privacy violations, and harm to fundamental rights. A structured AI risk assessment that addresses these dimensions — documented at the system level, linked to the inventory, reviewed on a defined cadence — satisfies the core risk documentation requirement across all three frameworks simultaneously.
Automated DPIA and privacy risk workflows are particularly relevant here because for AI systems processing personal data — which is most high-risk AI — the EU AI Act risk assessment and GDPR's Data Protection Impact Assessment requirement overlap directly. Article 26 of the EU AI Act explicitly directs deployers to use information provided by providers under Article 13 when conducting data protection impact assessments. A governance platform that integrates AI risk assessment with DPIA workflows eliminates the duplication that separate tracks create.
Transparency and documentation are shared obligations. All three frameworks require organizations to document their AI systems' purposes, capabilities, limitations, and data governance practices in ways that are accessible to relevant stakeholders. The AI RMF positions explainability as one of seven trustworthiness characteristics. ISO 42001 Annex A requires documentation of system purpose, components, data sources, intended and unintended impacts. The EU AI Act's Articles 13 and 14 require providers to produce instructions for use and for deployers to make transparency information available to affected persons. A unified documentation standard — one structured template covering all required disclosure elements — eliminates the redundant documentation maintenance that parallel tracks require.
Human oversight is a convergence point that goes beyond documentation. The EU AI Act's Article 14 requires high-risk AI systems to be designed to allow human intervention. The AI RMF's Govern function requires that human roles in AI decision-making are clearly defined. ISO 42001 requires named individuals with documented authority to intervene in live AI systems — including the power to pause, stop, or amend them. These three requirements point to the same operational control: a defined human oversight mechanism, assigned to named individuals, with documented escalation paths and the authority to act on AI system outputs. Designed once as an operational governance control, it satisfies all three.
Vendor and third-party AI risk management is another shared domain. The EU AI Act creates obligations throughout the AI supply chain — importers and distributors bear verification obligations, and deployers cannot disclaim responsibility for the behavior of systems they deploy. The AI RMF's Manage function includes supply chain risk management as a core governance activity. ISO 42001 Annex A addresses the governance of third-party AI relationships. For organizations whose AI estate includes third-party models, APIs, and embedded AI features — which as of 2025 describes approximately 78% of organizations — a unified vendor AI risk assessment process, with standard contractual provisions allocating compliance responsibilities, serves all three frameworks from a single governance workstream.
Incident response and post-market monitoring are shared requirements with slightly different trigger thresholds. The EU AI Act requires providers to report serious incidents to national authorities under Article 73 and requires deployers to maintain logs for six months. ISO 42001 requires continuous improvement triggered by incidents and nonconformities. The AI RMF's Manage function includes incident response planning. A single AI incident management procedure — covering detection, investigation, escalation, regulatory notification where required, and post-incident improvement — satisfies all three frameworks' incident governance requirements.
What Genuinely Requires Separate Treatment
Unified governance architecture does not mean identical governance for all purposes. Several requirements are genuinely framework-specific and must be addressed separately.
The EU AI Act has no equivalent in either the AI RMF or ISO 42001: the conformity assessment. Providers of certain high-risk AI systems under Annex III must complete a self-certification conformity assessment process before placing a system on the market — reviewing the system against the requirements of Articles 9 through 15, producing a conformity declaration, affixing CE marking, and registering the system in the EU AI database. For systems in sectors requiring third-party conformity assessment (a smaller subset), a notified body must be engaged. This is a regulatory gate that has no management-system equivalent; it must be completed as a distinct compliance activity. The technical documentation package required for conformity assessment — detailed technical specification, training data governance records, test results, monitoring plan — also exceeds what either the AI RMF or ISO 42001 require in their documentation provisions, and must be maintained in a format aligned with Annex IV of the regulation.
Prohibited AI practices under the EU AI Act — cognitive behavioral manipulation, social scoring, real-time biometric identification in public spaces with limited exceptions, emotion recognition in workplaces and educational settings, and others that became enforceable in February 2025 — require a categorical policy decision, not just a governance control. No ISO 42001 control and no AI RMF subcategory fully substitutes for the legal analysis of whether a system you operate or are developing falls within a prohibited category. That determination requires legal counsel, informed by the European Commission's published guidelines on prohibited AI practices, and must be documented in writing before any deployment decision. Learn more about EU AI Act Article 5 prohibited practices and grey areas.
ISO 42001 certification has a specific procedural requirement that neither the EU AI Act nor the AI RMF imposes: the formal management system infrastructure. This means management review meetings at defined intervals, a documented continual improvement cycle, internal audits against the standard's clauses, formal corrective action procedures, and a certification body relationship with surveillance audits. These are not governance activities that produce artifacts for other frameworks — they are specifically ISO process requirements that demonstrate the management system is functioning as designed. Organizations pursuing certification cannot substitute good governance outputs for this procedural layer; auditors check that the processes exist and are being followed, not just that the outputs are present.
The NIST AI RMF, by design, is more flexible than either of the other two frameworks on implementation specifics. It does not prescribe documentation formats, governance structures, or specific controls — it provides a vocabulary and a set of suggested practices that organizations adapt to their context. This flexibility is valuable for organizations at early stages of AI governance maturity, but it means the AI RMF alone does not provide the implementation specificity that either the EU AI Act's mandatory requirements or ISO 42001's Annex A controls provide. Organizations that adopt the AI RMF as their primary governance vocabulary should map its functions and subcategories to both ISO 42001 controls and EU AI Act requirements explicitly — using NIST's published crosswalk as the starting point — rather than assuming that AI RMF alignment implies compliance with the other two.
Building One Unified AI Governance Operating Model
The architecture of a unified governance program has five structural layers, each of which supports all three frameworks simultaneously.
This lifecycle is the architecture that holds all three frameworks together — and understanding how AI governance works across the AI lifecycle is what allows you to map EU AI Act articles, NIST RMF functions, and ISO 42001 controls onto a single program rather than three parallel workstreams.
Layer 1: AI Inventory and Classification
Everything else depends on knowing what AI systems you operate. The inventory must capture, for each system: the system's intended purpose and actual use; the data it processes and its sources; the populations it affects; the decisions or outputs it produces; who built it (internal, third-party, or combination); and its preliminary risk classification against both the EU AI Act's risk tiers and the organization's internal risk matrix. Systems that may fall under Annex III high-risk categories must be flagged for deeper assessment. Systems processing personal data must be linked to DPIA workflows.
The inventory is a living document, not a one-time audit output. It must be updated when new AI systems are deployed, when existing systems are significantly modified, and when new AI capabilities are integrated through third-party APIs or product updates. Shadow AI — models and AI-enabled tools adopted without central IT or governance review — is the most common inventory failure mode, and the one most likely to produce enforcement exposure under the EU AI Act's broad scope.
Layer 2: Unified Risk Assessment
A single AI risk assessment template, structured to capture the dimensions required by the EU AI Act (Article 9), ISO 42001 (Clause 6.1), and the AI RMF (Measure function), allows organizations to conduct one assessment per system and produce documentation that satisfies all three frameworks. The template should address: system purpose and context; identified risks across the dimensions of accuracy, bias, security, privacy, transparency, and fundamental rights; likelihood and impact scoring; existing controls and their effectiveness; residual risk determination; and the review trigger conditions that will require reassessment.
For AI systems processing personal data — which is the majority of high-risk AI — the risk assessment must be integrated with the DPIA process. The EU AI Act's Article 26 explicitly creates this linkage for deployers; building it into your unified risk assessment template ensures that one documented assessment satisfies both the AI Act's risk management obligation and GDPR's impact assessment requirement. This is one of the highest-value integration points in the entire unified architecture. The governance workflow tools available through privacy and compliance platforms increasingly support this integration by linking AI inventory records, risk assessments, and DPIA workflows within a single evidence system.
Layer 3: Policy Architecture
A unified AI policy architecture consists of a small number of high-level policies — acceptable use, AI lifecycle governance, data governance for AI, vendor AI risk management, and incident response — that are explicitly mapped to the provisions of the EU AI Act, ISO 42001 Annex A controls, and AI RMF governance subcategories. Mapping these references into policy documents directly (a policy section can note which Article or Clause it implements) creates an evidence trail that dramatically simplifies audit preparation across all three frameworks. Learn more about acceptable use policies covering Article 5 prohibited categories.
The policy architecture must be owned cross-functionally. Legal owns the EU AI Act compliance mapping. Privacy owns the data governance provisions. Security owns the risk management and incident response policies. Product and engineering own the AI lifecycle governance provisions. Each owner must understand which governance obligations they are implementing — not just the organizational policy they are following — so that policy updates triggered by regulatory change are routed to the right people.
Layer 4: Evidence Management
Audit readiness for all three frameworks depends on evidence that is current, traceable, and retrievable. The EU AI Act's technical documentation must reflect the actual current state of deployed systems, not a historical snapshot. ISO 42001 surveillance audits will require evidence that management reviews occurred, internal audits were conducted, corrective actions were completed. AI RMF alignment assessments require evidence that governance practices are operational, not just documented in policies.
The practical implication is that evidence management must be a continuous operational function, not a pre-audit sprint. Privacy governance platforms that centralize records of processing activities, risk assessments, impact assessments, and vendor assessments provide the organizational infrastructure that makes this continuous evidence management feasible at scale — maintaining the living documentation system that all three frameworks require rather than treating compliance documentation as a periodic project deliverable.
Layer 5: Ongoing Monitoring and Improvement
All three frameworks require governance that responds to change — in AI system behavior, in regulatory requirements, and in the organization's AI estate. The EU AI Act's post-market monitoring obligation requires providers to track system performance and report serious incidents. ISO 42001's continual improvement model requires that nonconformities trigger documented corrective action. The AI RMF's iterative structure assumes that risk management cycles repeat as systems and contexts change.
Model drift — where an AI system's behavior shifts after deployment because the input data distribution has changed — is the most common operational trigger for re-assessment. A system that was assessed and documented at deployment may be materially different after several months of production operation, particularly if it involves continuous learning or is regularly retrained. Organizations must build governance hooks into their ML operations pipelines so that material model changes trigger documentation updates, risk reassessment, and where applicable, updated conformity assessment.
Implementation Sequencing for Enterprise Organizations
The most common implementation failure is attempting to achieve compliance with all three frameworks simultaneously from scratch. A more effective sequence builds governance maturity in layers that each create value while laying infrastructure for the next.
Start with the AI inventory and high-risk classification. This is the foundation of everything else and has immediate enforcement relevance — organizations that cannot classify their AI systems under the EU AI Act's risk tiers cannot complete conformity assessments, cannot establish meaningful risk management records, and cannot structure their governance program appropriately. The inventory also enables the vendor AI risk assessment process that all three frameworks require.
Build the unified risk assessment process second. Using NIST's published AI RMF-to-ISO 42001 crosswalk as a mapping reference, design a risk assessment template that captures the required elements of all three frameworks in a single document. Integrate it with your DPIA process for AI systems processing personal data. Conduct initial assessments for all systems in the inventory, prioritizing potential Annex III high-risk systems.
Establish the policy architecture and ownership model third. Clear cross-functional accountability — documented in writing, with named owners for each governance domain — is what makes a governance program sustainable rather than dependent on individual knowledge. The policy architecture also provides the "management system" layer that ISO 42001 requires: policies that are formally approved, reviewed on a defined cycle, and mapped to the framework requirements they implement.
Address EU AI Act conformity assessment obligations for high-risk systems as a fourth workstream. These are EU AI Act-specific requirements that require dedicated attention: completing the Annex IV technical documentation package, conducting the conformity assessment process, drawing up the EU declaration of conformity, and registering systems in the EU database. The EU AI Act compliance guide for enterprises covers this workstream in detail, including the documentation architecture that regulators expect.
Pursue ISO 42001 certification fifth, once the governance infrastructure is in place. The management system processes — management reviews, internal audits, corrective action procedures — are easier to establish and evidence when the substantive governance content (inventory, risk assessments, policies, monitoring) already exists. Organizations that attempt to pursue certification before building governance substance end up with well-documented but operationally empty management systems that fail surveillance audits.
Frequently Asked Questions
What is the difference between the EU AI Act and NIST AI RMF?
The EU AI Act is binding law with extraterritorial scope, risk-based obligations, and financial penalties up to 7% of global annual turnover. The NIST AI RMF is a voluntary US framework with no enforcement mechanism. The AI Act imposes specific mandatory controls on high-risk AI providers and deployers; NIST provides a flexible governance vocabulary that organizations adapt to their context. Many organizations use the AI RMF as their operational governance language while ensuring that language maps to the specific compliance obligations of the EU AI Act.
Is ISO 42001 required for AI compliance?
ISO 42001 certification is not legally required by any regulation. However, it is increasingly required by enterprise procurement processes and RFPs, and it provides documentary evidence relevant to the EU AI Act's Article 17 quality management system requirement. Organizations with existing ISO 27001 infrastructure can achieve ISO 42001 certification with significantly reduced implementation effort because the management system structure is shared.
Can one governance program satisfy all three frameworks?
Yes, if designed correctly. The substantial overlap in foundational requirements — AI inventory, risk assessment, documentation, human oversight, vendor management, incident response — can be addressed through a single set of controls mapped to all three frameworks simultaneously. What cannot be unified are framework-specific procedural requirements: EU AI Act conformity assessments, ISO 42001 management system processes, and the EU Act's specific prohibited AI determinations each require dedicated treatment.
Is the NIST AI RMF legally binding?
No. The NIST AI RMF is explicitly voluntary and carries no enforcement penalties. However, in regulated sectors in the US, NIST AI RMF alignment is increasingly referenced in regulatory guidance and federal procurement requirements. Not adopting it creates reputational and procurement risk even without legal compulsion.
Does ISO 42001 help with EU AI Act compliance?
Significantly. ISO 42001 certification provides a structured management system framework that aligns with the EU AI Act's Article 17 quality management system requirement. The standard's risk assessment, documentation, and continuous improvement requirements overlap substantially with the EU AI Act's high-risk AI obligations. While certification does not substitute for conformity assessment, it demonstrates systematic AI governance that regulators treat as evidence of organizational seriousness about compliance.
How should multinational companies approach AI governance?
Multinationals should build governance around the most demanding framework they face and map downward. For most multinationals with EU market exposure, that means building the EU AI Act compliance architecture — which is the most prescriptive — and demonstrating explicitly how that architecture satisfies ISO 42001 requirements and NIST AI RMF functions. This approach avoids building up to each framework separately and produces governance documentation that satisfies all requirements from a unified evidence base.
What teams should own AI governance programs?
Effective AI governance is inherently cross-functional and cannot be owned by any single team. Legal and compliance own the EU AI Act regulatory mapping. Privacy owns data governance for AI systems processing personal data. Security owns risk assessment and incident response. Product and engineering own AI lifecycle governance and the technical documentation of deployed systems. Vendor management owns third-party AI risk. The governance program needs a named executive owner and a cross-functional steering committee — not because governance is complicated, but because it touches every part of the organization that builds, buys, or deploys AI.
Conclusion: One Program, Not Three
The organizations that will emerge from the current wave of AI governance requirements with sustainable, scalable compliance programs are not those that have built the most elaborate parallel compliance tracks. They are those that recognized early that the EU AI Act, NIST AI RMF, and ISO/IEC 42001 are three instruments pointing at the same underlying obligation — governed, accountable, documented AI — and built a single program to address all three.
The unified architecture is not a shortcut. It requires the same foundational governance work that any one of these frameworks demands: an accurate AI inventory, system-level risk assessment, cross-functional accountability, continuous documentation, and operational monitoring. What it eliminates is the duplicated effort, inconsistent documentation, and audit fatigue of running three disconnected compliance programs across teams that should be working from the same evidence base.
The August 2026 EU AI Act enforcement deadline is the current forcing function. But the governance infrastructure that date demands — inventory, risk management, documentation, human oversight, vendor controls, incident response — is also the infrastructure that ISO 42001 auditors will examine and that AI RMF alignment assessments will measure. Build it once, map it to all three, and maintain it as a continuous operational program rather than a periodic compliance project.
Assess your AI governance maturity against all three frameworks. Map your existing controls to the unified control domains. Identify the gaps — conformity assessments pending, vendor contracts missing AI governance provisions, documentation not current. That gap analysis is the implementation roadmap.
