Why the next wave of AI value will come from connecting strategy, capabilities, value streams, governance, and delivery
By Paul Gadbois
AI has moved beyond the experimentation stage. The executive question is no longer, “Should we use AI?” It is, “How do we make AI produce measurable enterprise value without creating more fragmentation, risk, and operational noise?”
That question cannot be answered by tooling alone. It cannot be solved by a center of excellence that sits apart from delivery. And it certainly cannot be solved by letting every function pursue its own AI agenda independently.
The organizations that will create durable value from AI are the ones that treat AI as an enterprise operating model challenge, not just a technology adoption challenge.
This is where enterprise architecture must step forward.
Gartner’s 2026 strategic technology trends highlight AI-native development platforms, multiagent systems, domain-specific language models, digital provenance, AI security platforms, and governance-oriented foundations as major forces shaping enterprise strategy. Gartner also frames these trends around the need to align digital strategy with enterprise goals, scale AI securely and responsibly, and protect enterprise value [1].
That is not a narrow technology agenda. It is an operating model agenda.
The AI problem is becoming an execution problem
Many organizations now have AI strategies. Fewer have an AI operating model.
An AI strategy may identify use cases, platforms, vendors, and investment themes. An AI operating model defines how the enterprise will repeatedly convert those ideas into governed, measurable, scalable business outcomes.
That distinction matters.
Without an operating model, AI often becomes a collection of disconnected experiments. Marketing builds content automation. Finance pilots forecasting. Operations explores agents. Technology teams evaluate platforms. Risk teams develop policies after the fact. Everyone is busy, but the enterprise may not be getting meaningfully better.
The work becomes fragmented because the connective tissue is missing.
Enterprise architecture is uniquely positioned to provide that connective tissue because it already sits at the intersection of strategy, business capabilities, value streams, information, applications, platforms, governance, and delivery. Forrester’s 2026 enterprise architecture overview describes EA as critical for enabling organizations to scale, operate, and deliver technology strategies and services, aligning business goals with IT infrastructure and systems across people, process, information, and technology [2].
That is exactly the discipline required for AI at enterprise scale.
The modern enterprise architect must become a value orchestrator
The Modern Enterprise Architect is not merely a framework custodian, solution reviewer, or standards enforcer. Those activities still matter, but they are not enough.
In the AI era, the modern enterprise architect must help answer five executive questions:
- Which business outcomes are we trying to improve?
- Which capabilities must change to achieve those outcomes?
- Which value streams will be redesigned or accelerated?
- Which data, technology, and AI services are required?
- How will we govern value, risk, adoption, and delivery together?
This is a shift from architecture as documentation to architecture as value orchestration.
The architect’s job is not to slow the organization down. It is to make change more coherent. It is to help the enterprise move faster in the right direction, with fewer disconnected investments and fewer avoidable risks.
Start with capabilities, not tools
AI conversations often begin with tools: copilots, agents, models, automation platforms, data products, and security services. Those tools may be important, but they should not be the starting point.
The better starting point is the business capability model.
A capability model gives executives and architects a stable view of what the enterprise must be able to do, independent of current org charts, projects, or applications. It creates a practical bridge between strategy and execution.
For example, if the enterprise strategy calls for improved customer retention, AI investments might target capabilities such as customer insight, next-best-action recommendation, service case resolution, churn prediction, knowledge management, or personalized engagement.
The question becomes: Which capabilities must perform differently for the strategy to succeed?
That question is more useful than: Where can we deploy AI?
This is where The Goal GPS™ Framework becomes valuable. Goals define the destination. Performance measures define whether progress is real. Strategy defines the path. Capabilities define what must be strengthened. Value streams reveal where work actually flows. Technology enables the change. Governance keeps the system aligned.
Value streams reveal where AI can create measurable impact
Capabilities tell us what the organization must be able to do. Value streams show how value is delivered across functions.
AI value is often lost because organizations optimize tasks rather than outcomes. A team may automate a document review, summarize customer calls, or generate code faster. Those improvements may be useful, but they do not automatically improve enterprise performance.
The stronger question is: Where does the value stream slow down, break down, create rework, introduce risk, or fail the customer?
A value-stream lens helps identify AI opportunities that matter:
- Where are decisions delayed?
- Where is knowledge trapped?
- Where are handoffs excessive?
- Where is manual review necessary but inconsistent?
- Where does poor data quality create downstream cost?
- Where could AI augment human judgment rather than replace it?
This reframes AI from isolated productivity experiments to enterprise delivery improvement.
Governance must become embedded, not episodic
AI governance cannot operate as a late-stage approval board. That model will not scale.
Governance must be embedded into the way AI-enabled capabilities are designed, funded, built, deployed, monitored, and retired.
NIST’s AI Risk Management Framework is designed to help organizations manage AI risks and incorporate trustworthiness considerations into the design, development, use, and evaluation of AI systems. NIST also released a Generative AI Profile in 2024 to help organizations identify generative AI risks and actions aligned to their goals and priorities [3].
For enterprise architects, this means AI governance should be translated into architecture decisions and delivery practices:
- Model and agent inventories.
- Approved patterns for human-in-the-loop control.
- Data classification and usage boundaries.
- Traceability between AI use cases and business capabilities.
- Risk tiers based on business impact.
- Security, privacy, and compliance controls built into platforms.
- Monitoring for quality, drift, adoption, cost, and business value.
- Clear ownership for outcomes and exceptions.
The goal is not governance for its own sake. The goal is confidence at scale.
The operating model must connect funding, delivery, and accountability
A recurring reason that transformation efforts underperform is that strategy, funding, architecture, and delivery operate on different clocks.
Strategy is annual. Funding is often portfolio-based. Architecture may be project-triggered. Delivery is agile or product-oriented. Governance is periodic. Metrics are inconsistent.
AI will amplify the weakness of that model.
Accenture’s 2025 work on modern IT operating models recommends connecting every point of IT strategy to the business agenda, creating clear metrics that link technology initiatives to business outcomes, identifying where AI and emerging technologies can deliver value, doubling down on product-oriented teams, implementing secure technology platforms, and exploring agentic AI in enterprise architecture [4].
That guidance points to a practical operating model principle: AI should be governed and delivered through business outcomes, not isolated technology projects.
A mature AI operating model should connect:
- Strategic objectives.
- Critical success factors.
- Business capability gaps.
- Value-stream improvement opportunities.
- AI and data product roadmaps.
- Enterprise architecture standards and patterns.
- Funding and portfolio decisions.
- Delivery teams and platform teams.
- Risk, compliance, and security requirements.
- Adoption and value realization metrics.
This is not bureaucracy. It is the management system required to convert AI ambition into enterprise results.
A practical blueprint for AI operating model design
Executives do not need another abstract maturity model. They need a way to organize the work.
A practical AI operating model can be designed around seven components.
Strategic intent
Define where AI matters to the business. Tie AI themes directly to enterprise goals, critical success factors, and measurable outcomes.
Avoid generic aspirations such as “become AI-enabled.” Instead, define the business performance that must improve.
Capability focus
Map AI opportunities to business capabilities. Identify which capabilities need augmentation, automation, acceleration, or redesign.
This prevents tool-led investment and keeps the conversation grounded in what the enterprise must be able to do.
Value-stream redesign
Use value streams to find friction, delay, rework, risk, and decision latency. Prioritize AI use cases that improve flow and measurable outcomes.
This helps separate interesting use cases from important ones.
Data and technology foundations
Define the enabling architecture: data products, integration patterns, AI platforms, model management, security services, identity, observability, and controls.
This is where architecture turns ambition into reusable enterprise capability.
Governance and risk management
Embed governance into intake, design, delivery, deployment, monitoring, and continuous improvement. Use risk tiers so governance is proportional to business impact.
Not every AI use case needs the same level of oversight. But every AI use case needs clear accountability.
Product and platform delivery
Organize delivery around durable products, platforms, and capability teams where possible. AI should not become a side project trapped in innovation theater.
Platform teams should make the right patterns easier to use than the wrong ones.
Value realization
Track adoption, business outcomes, risk indicators, operational performance, and cost. Measure value after deployment, not just at business-case approval.
AI value realization must be managed as an ongoing discipline.
The Enterprise Architect’s role in the AI operating model
Enterprise architects should not try to own AI alone. That would be a mistake.
AI operating models require partnership across business leadership, technology, data, security, risk, legal, finance, HR, product, and operations.
But enterprise architects can play a central orchestration role by creating the shared structure for decision-making.
The architect can help the enterprise answer:
- Which AI investments support strategic goals?
- Which capabilities and value streams are affected?
- Which architecture patterns should be reused?
- Which platforms should be standardized?
- Which risks require governance controls?
- Which delivery teams are accountable?
- Which metrics will prove value?
- Where are we creating duplication or unmanaged complexity?
This is where enterprise architecture becomes practical and executive-relevant.
Practical takeaways for leaders
- Do not start with an AI tool inventory. Start with enterprise goals, critical success factors, capabilities, and value streams.
- Make AI governance part of the delivery system. A committee can set direction, but embedded controls create scale.
- Create traceability from strategy to execution. Every AI investment should connect to a business outcome, capability, value stream, technology enablement, and measurable result.
- Treat AI platforms as enterprise capabilities. Reusable patterns, security controls, data foundations, and monitoring services will matter more than isolated experiments.
- Expand the role of enterprise architecture. The modern enterprise architect should help orchestrate value, manage complexity, guide technology alignment, and connect strategy to execution.
Closing thought
AI will not fix a disconnected operating model. It will expose it.
The enterprises that win with AI will not be the ones with the longest list of pilots. They will be the ones that build the management system to turn strategy into governed, measurable, repeatable execution.
That is the real work that enterprise architecture was meant to do.
About the Author
Paul Gadbois is a technology executive, enterprise architect, author, and advisor focused on helping organizations align strategy, business capabilities, and enabling technology. He is the author of The Modern Enterprise Architect and creator of The Goal GPS™ Framework. Learn more at paulgadbois.com.
Sources Cited
[1] Gartner. “Gartner Top 10 Strategic Technology Trends for 2026.”
[2] Forrester. “Enterprise Architecture Overview / Enterprise Architecture Services,” 2026.
[3] National Institute of Standards and Technology (NIST). “Artificial Intelligence Risk Management Framework (AI RMF 1.0)” and “Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile,” 2024.
[4] Accenture. “Modern IT Operating Model,” 2025.