Why AI Strategy Belongs in the COO's Office
95% of enterprise AI pilots fail to produce measurable financial returns.
This isn't a technology gap. It is a structural failure. As enterprises pour billions into AI, 70% to 90% of projects die in pilot purgatory, never scaling beyond the lab. The root cause isn't the algorithm; it's the org chart. Most AI initiatives are trapped under IT ownership.
This is an organizational architecture challenge. When AI lives under the COO, it stops being a cost center and becomes the operating system.
The Incentive Misalignment: Why IT-Led AI Fails
Putting AI under the CIO isn’t about politics, it’s about incentives. IT is built to stabilize systems; AI transformation is built to disrupt them. Those two missions don’t mix.
An IT-Led Mandate is (rightfully) focused on stability, security, and cost efficiency. Its KPIs are technical: system uptime, availability, and IT cost reduction. Success is a stable, secure platform.
An Operations-Led Mandate is focused on performance, margin, and speed. Its KPIs are business outcomes: cycle time reduction, margin improvement, and customer satisfaction.
AI transformation isn’t an IT upgrade; it’s an operational redesign. IT is paid to keep things stable. AI breaks stability. That’s why IT-led programs chase technology, while operations-led ones chase results.
The Integrator Mandate: Why the COO Must Lead
The modern COO is no longer just an execution steward; they're the architect of transformation. They sit at the intersection of every function, uniquely positioned to weave AI into the fabric of daily operations, turn data into decisions, and lead transformation across the enterprise.
The numbers make the case pretty clearly. Gartner found that when AI or digital programs sit in silos, less than half hit their targets. When operations and IT actually share ownership, the success rate jumps to 71%. That shared, ops-led model is the clearest signal of real ROI.
The Evidence: Operations-Led AI Delivers Real ROI
Across manufacturing, insurance, and tech, a consistent pattern emerges: AI delivers returns only when operational leaders drive the redesign, not when IT departments simply deploy tools.
This performance delta isn't subtle. A European equipment manufacturer's COO led a centralized AI reassessment of core operating assumptions, producing a prioritized roadmap for an estimated €300 million in EBITDA improvement. An insurer that embedded AI into its claims process—an operational, not technical, project—saw a 40% reduction in claim handling time and a 15-point increase in NPS.
This contrasts sharply with the IT-led model, where an estimated 95% of GenAI pilots produce no measurable P&L impact and over 60% of IT-led AI projects stall. The structural lesson is clear: success is driven by managing organizational change and redesigning the workflow, not just by deploying a new tool.
Designing the Operating Model: How to Govern at Scale
An operations-led strategy needs a strong, scalable architecture. The hybrid hub and spoke model has proven the most effective because it provides a clear decision framework for scaling while balancing control with agility.
It begins as a centralized function to establish a solid foundation: a Center of Excellence (COE) reporting to the COO. The COE sets the governance framework, risk policies such as the NIST AI RMF, and a shared technology stack.
As capabilities mature, the model transitions to the optimal hybrid state:
The Central Hub (COO's Office) owns enterprise strategy, governance, ethical and risk frameworks, and platform standards. This guarantees control, security, and efficiency.
The Business Unit Spokes (P&L owners) own use case prioritization, business case and ROI, data stewardship, and frontline change management and adoption. This ensures speed, agility, and business relevance.
Structure alone isn't enough. Choosing the wrong metrics can still sink the initiative.
The New Scorecard: Measuring What Matters
A new operating model needs a new scorecard. We have to stop measuring AI like an IT utility and start measuring it like a business driver.
The key is to tie these new, ops-focused KPIs directly to executive scorecards to create real accountability.
Governing Autonomous Agents: The COO Imperative
The rise of Agentic AI makes this architectural shift urgent. These are autonomous agents that can reason, plan, and act across workflows. The operational implication is that the COO's role is shifting from managing processes to governing digital workers.
This isn't a futuristic fantasy; seven in ten COOs are already embracing this shift. Without an operational framework in place, these agents will create chaos faster than value. The COO must establish an operating framework for governing autonomous agents, managing their lifecycles, auditing their decisions, and designing human-in-the-loop escalation paths.
The Roadmap: Three Strategic Imperatives to Take Ownership
This is a practical guide for COOs to take control of the AI agenda through decisive, operations-led actions.
Assess the Baseline: Launch an enterprise readiness scorecard across data, governance, and culture. This establishes a factual baseline, identifies critical gaps, and builds the political case for change.
Build the Hub: Establish a central AI Center of Excellence (COE) that reports directly to the COO. This centralizes authority, standardizes governance, and shifts the focus to a high-ROI, operations-led portfolio.
Formalize the Charter: Solidify C-suite alignment on a formal AI charter. This document codifies the hub-and-spoke model, C-suite decision-rights, business-value KPIs, and the 12-month evolution plan.
The Baldrige Moment: A Mandate for Change
The 1980s quality crisis was an organizational crisis. The Malcolm Baldige National Quality Award was a systemic framework for reform, not a new piece of technology. The result was an 820-to-1 societal benefit-to-cost ratio.
Baldrige wasn't about better machines, it was about better management systems. AI demands the same shift.
AI won’t transform your company, your operating model will. The question is whether you’ll design it or inherit it.