Editorial
eBook
Web Marketing Moneyball: How Web Operations Drive Marketing ROI
The Gist
- AI readiness is not tool access. Models and copilots are easier to adopt, but production value still depends on data quality, integrations, workflow design and ownership.
- The AI stack is becoming the strategy. Compute, chips, data centers, energy, networks and security now shape what organizations can deploy and scale.
- Governance turns AI into capability. The winners will be organizations that can operationalize AI safely, repeatedly and measurably across real business workflows.
During IEEE-USA Congressional Visits Day, I joined technology professionals in a series of Capitol Hill discussions with congressional offices on federal R&D, AI competitiveness, STEM talent and America’s innovation capacity. A discussion with Senator Andy Kim’s office especially reinforced a point enterprise leaders should take seriously: AI leadership is often described as a model race, but in practice it is a systems race.
A systems race means the winner is not simply the organization with access to the strongest model. It is the organization with the infrastructure, data, governance, talent and workflows to deploy AI repeatedly and safely.
That point matters now because AI adoption has moved faster than enterprise readiness. Many organizations are experimenting with copilots, chatbots, knowledge assistants and generative AI tools. But experimentation is not the same as operational capability.
The public conversation often focuses on who has the best large language model, the most advanced chatbot or the next breakthrough release. Those things matter. But they are only one layer of the competitive equation.
The more durable advantage will come from what surrounds AI: infrastructure, proprietary data, integration architecture, governance, workflow redesign and the ability to move from impressive demonstrations into trusted production outcomes.
Recent research supports that shift. Stanford HAI’s AI Index 2025 found that GPT-3.5-level inference costs fell from $20 to $0.07 per million tokens between November 2022 and October 2024, while AI hardware price-performance improved about 30% annually and energy efficiency about 40% annually.
That is good news. But as AI becomes cheaper and more available, enterprise advantage moves toward the systems that make AI useful, safe and trusted inside real customer, employee and operational workflows.
Table of Contents
- The AI Race Is Moving Beyond Models
- Infrastructure Is Now Part of AI Strategy
- Tool Access Is Not AI Readiness
- Customer-Facing AI Shows Why Systems Matter
- Governance Turns AI Into Operating Capability
- An AI Systems Readiness Checklist
- The Model Race Shapes Headlines. The Systems Race Creates Value.
The AI Race Is Moving Beyond Models
Models are the most visible part of AI, but they are not the whole stack.
A model depends on compute infrastructure, data pipelines, integration architecture, cybersecurity, energy availability, domain knowledge, workflow design and human oversight. In many industries, it also depends on physical systems: sensors, robotics, chips, networks and components that connect intelligence to the real world.
OECD research on AI infrastructure describes frontier AI as a supply chain that depends on accelerator chips, advanced networking, foundries, memory, electronic design automation tools and hyperscale cloud capacity. It also notes that the three largest cloud providers hold more than 60% of the global cloud market.
For enterprise leaders, AI competitiveness is not only happening in model labs. It is happening in cloud contracts, data-center capacity, chip supply chains, network architecture, system integrations and data governance.
This becomes especially important as AI moves into healthcare, manufacturing, logistics, financial services, retail operations and customer experience. In those environments, the question is not only whether AI can produce a fluent answer. The question is whether the organization can safely embed that answer into a decision, process or customer interaction.
An enterprise may have access to a capable model but still lack clean data, reliable systems of record, clear escalation logic, secure integrations, trained users or measurable business outcomes. In that case, the model is not the constraint. The operating environment is.
Related Article: AI Broke Your Content Governance — Now What?
Infrastructure Is Now Part of AI Strategy
The systems race is not just a metaphor. It is becoming physical.
As AI scales, it requires more data-center capacity, power, cooling, specialized chips, networking and capital. AI strategy is becoming inseparable from infrastructure strategy.
The International Energy Agency projects global data-center electricity demand to nearly double from 485 TWh in 2025 to 950 TWh in 2030, while electricity consumption from AI-focused data centers is expected to triple over the same period. The 2024 Lawrence Berkeley National Laboratory U.S. data-center report estimates that U.S. data centers used 176 TWh in 2023 and could reach 325 to 580 TWh by 2028, or 6.7% to 12% of total U.S. electricity consumption.
For most enterprise leaders, this does not mean becoming an energy expert. It does mean AI infrastructure can no longer be treated as a back-office IT matter.
The availability, cost and resilience of compute, cloud, power and network capacity will increasingly shape what organizations can deploy, where they can deploy it and how fast they can scale.
This is also why federal R&D and long-term technology investment matter. The technologies companies depend on commercially often start upstream in research, standards, infrastructure and public-private investment. AI, advanced communications, cybersecurity, robotics, photonics and intelligent infrastructure will not scale on model capability alone.
Tool Access Is Not AI Readiness
Inside companies, the biggest gap is no longer access to AI. It is the ability to turn access into measurable enterprise value.
McKinsey’s 2025 State of AI survey found that nearly nine in 10 respondents report regular AI use in their organizations, but only 39% report enterprise-level EBIT impact and only about 6% qualify as AI high performers.
That is the readiness gap in one statistic.
Many organizations have moved quickly to give employees access to AI tools. But access does not mean the organization has redesigned workflows, improved data quality, clarified ownership, trained teams, governed risk or measured impact.
BCG found that 74% of companies still struggle to achieve and scale value from AI, while only 26% have built the capabilities needed to move beyond proofs of concept. Gartner research has also pointed to recurring AI project failures after proof of concept because of poor data quality, weak risk controls, escalating costs or unclear business value.
Webinar
Jun
16
From Legacy to Launch-Ready: How Gainbridge Made Its Website a Marketing-Led Growth Engine
Join in to learn how a D2C annuity brand gave marketing full website ownership — without slowing down or risking compliance.
Register
Webinar
Jun
18
The 5-Question CX Audit: Benchmark Your CX Operations for 2026
Built around insights from the 2026 CX Outsourcing Report, this live session puts the audit into practice.
Register
Webinar
Jun
30
How Modern Marketing Is Exposing the Limits of Legacy CMS
Why marketers are rethinking CMS workflows that slow publishing, personalization and campaign execution.
Register
Webinar
Jul
14
Content Leaders Collective: Proving Content’s Business Impact
Join us as top content leaders look beyond the buzzwords to share how they actually prove ROI and scale what works.
Register
Webinar
On demand
Content Strategy Leaders Live: Managing Risk, Compliance & AI in Financial Services
Learn how financial services leaders are modernizing content systems without disrupting trust, compliance or experience.
Watch Now
Webinar
On demand
From AI Investment to CX Results: What Enterprise Leaders Need to Know
Move beyond experiments. See how top enterprises scale AI for CX results.
Watch Now
This is where many enterprise AI strategies become too narrow. They focus on the tool layer while underinvesting in the readiness layer.
A pilot can look impressive because it is tested in a controlled environment. Production is different. Production has messy data, legacy systems, cross-functional ownership, real users, security reviews, legal concerns, customer expectations and operational accountability.
The next AI divide will be between organizations experimenting with AI and organizations that can operationalize AI as a repeatable capability.
Customer-Facing AI Shows Why Systems Matter
Customer-facing AI is one of the fastest ways to reveal whether an enterprise is actually ready.
When AI is used internally, mistakes may stay contained. When AI enters the customer journey, errors become visible immediately. A wrong product recommendation, inaccurate order update, broken chatbot handoff or misleading promotion can damage trust quickly.
This is especially true in high-consideration journeys. Customers do not only want answers. They want confidence. They want to know that the information is accurate, the offer is valid, the status is real and help is available when the journey becomes complex.
That means AI in customer experience needs more than conversational fluency. It needs operational truth.
In enterprise digital experience, the hardest part is rarely the demo. It is connecting the demo to pricing, inventory, identity, fulfillment, service, compliance and measurement.
If an AI assistant tells a customer that an order is ready, the fulfillment system must agree. If it recommends a product, inventory and eligibility rules must support the recommendation. If it offers a promotion, pricing and checkout systems must recognize it. If it escalates to a human, the handoff must preserve context.
Without that alignment, AI becomes another promise the enterprise cannot keep.
This is not only a CX issue. It is a systems issue. The same pattern appears in employee workflows, partner portals, service operations, supply chain decisions and internal automation.
Actions CX Leaders Should Take Before Scaling Customer-Facing AI
Editor's note: AI success is no longer determined by model access alone. Customer experience leaders increasingly need the data, governance, systems integration and operational discipline required to turn AI pilots into trusted customer outcomes. This table highlights practical actions CX leaders can take to improve AI readiness and reduce deployment risk.
| AI Readiness Area | Key Question for CX Leaders | Recommended Action | Expected CX Impact |
|---|---|---|---|
| Customer Data Quality | Can AI access accurate, current customer information? | Audit customer data sources, eliminate duplicates and establish ownership for data quality. | Improves personalization accuracy and reduces customer frustration. |
| Journey Integration | Is AI connected to the systems that support customer interactions? | Integrate AI with CRM, commerce, service and fulfillment platforms before expanding use cases. | Creates consistent, reliable customer experiences. |
| Governance & Oversight | Are there clear rules for when AI acts independently versus escalating? | Define approval thresholds, escalation paths and human-review requirements. | Reduces risk while maintaining customer trust. |
| Operational Truth | Can AI verify information before presenting it to customers? | Ground AI responses in approved systems of record and real-time operational data. | Prevents inaccurate recommendations, status updates and promotions. |
| Employee Enablement | Do frontline teams understand how AI supports customer journeys? | Train service, marketing and digital teams on AI workflows, limitations and escalation procedures. | Improves adoption and customer issue resolution. |
| Workflow Design | Is AI embedded within a measurable business process? | Assign process owners and establish KPIs before deployment. | Moves AI from experimentation to business value. |
| Customer Trust | Can customers easily reach a human when needed? | Design seamless handoffs that preserve customer context and conversation history. | Protects customer satisfaction during complex interactions. |
| Performance Measurement | How will success be measured? | Track containment, resolution, satisfaction, revenue impact and operational efficiency metrics. | Provides evidence of business value and customer impact. |
| Risk Management | What happens when AI makes a mistake? | Create monitoring, incident response and rollback procedures. | Minimizes disruption and reputational damage. |
| Scaling Strategy | Which use cases deserve broader investment? | Prioritize high-value customer journeys with measurable outcomes before expanding. | Accelerates ROI while reducing implementation risk. |
Governance Turns AI Into Operating Capability
Governance is often treated as a brake on innovation. That is the wrong framing.
Good governance is what allows AI to scale.
It defines where AI can act independently, where it must cite a source, where it must ask for approval, where it must escalate to a human and how the organization will measure quality. It also clarifies accountability when the system fails.
NIST’s Generative AI Profile frames AI risk management as a lifecycle discipline, including provenance, monitoring, human oversight, third-party dependency, incident response and operational fallback plans.
That is the mindset enterprises need.
AI governance should not start after a bad answer reaches a customer, employee or business partner. It should be part of how use cases are selected, designed, tested, deployed and monitored.
The goal is not to slow down every AI initiative. The goal is to make sure AI can survive production reality.
The strongest organizations will not be the ones with the most pilots. They will be the ones that know which AI use cases deserve production investment, which require human oversight and which should not be automated at all.
An AI Systems Readiness Checklist
Before scaling AI, leaders should ask whether the model is connected to the data, systems, governance, people and workflow needed to support the outcome it promises.
| Readiness area | What leaders should verify |
|---|---|
| Data | Is the model grounded in accurate, current and approved data sources? |
| Systems | Is AI connected to the systems of record needed to support the answer, decision or action? |
| Governance | Are boundaries clear for when AI can act, when it needs approval and when it must escalate? |
| People | Do the teams using or supervising AI understand the workflow, risks and success metrics? |
| Workflow | Is the AI embedded in a real business process with ownership, measurement and feedback loops? |
This checklist helps shift AI strategy from “Can the model respond?” to “Can the enterprise support the outcome?”
The Model Race Shapes Headlines. The Systems Race Creates Value.
AI leadership will not be won by prompts alone. It will not be won by access alone. And it will not be won by experimentation alone.
The next phase of AI competition will be won by the companies, regions and countries that can connect research, talent, infrastructure, governance and execution into one repeatable engine.
For enterprise leaders, that means moving toward AI operating maturity. It means building the systems that let AI work inside real business processes, customer journeys and employee workflows.
The model race matters.
But the systems race will determine who creates lasting value.
Learn how you can join our contributor community.