We need a better leadership operating system to create enterprise value in the AI era
Writing: Peter Whealy
Over three-quarters of AI transformations fail to deliver expected value. Technology is advancing at an extraordinary speed, yet the leadership architecture governing it is not.
Technological revolutions have always exposed organizational lag. Early factories installed machines but kept management built for manual workshops. Railways expanded before anyone designed a system to run them safely. In both cases, the technology outpaced the thinking around it.
The same pattern is once again emerging today. AI is being deployed at scale, yet leadership systems designed for hierarchy, siloed expertise and sequential execution remain intact. That structural mismatch is exactly where enterprise value leaks. AI does not fail because of technology immaturity, but because organizations attempt to run a new intelligence architecture on a leadership system designed for a slower, more predictable world. As a result, most organizations deploy AI in isolated pockets, aiming to improve individual functions without redesigning an integrated system.
AI can accelerate individual functions, but until the leadership operating system is rebuilt for the AI era, acceleration will compound fragmentation rather than resolve it. Our challenge is not AI adoption. It is redesigning the leadership system so that orchestration becomes possible.
Three tensions defining the leadership crisis
In researching my book Lead with AI. Stay Human., I conducted more than 50 interviews with senior executives across global organizations. Three tensions surfaced consistently regardless of industry, geography or seniority. They are structural realities shaping enterprise performance today.
The identity tension
Senior leaders have built their careers through credibility and mastery – being the person with the answers, the one who applied experience under pressure and brought clarity to complexity. Then AI democratized those core abilities.
Ernst Wühr, CEO of EMS Medical Devices, describes the shift: “My role has shifted from primary decision-maker to the orchestrator of humans and AI intelligence.”
What has emerged is not a skills gap but a credibility disruption. It is common across industries, yet it is rarely discussed explicitly inside organizations. Leaders who cling to behaviors that once signaled competence now signal constraint. Those who insist on being the final verifier become a bottleneck – and as AI systems improve, that bottleneck becomes more visible and more costly. Decisions begin to stall and transformation slows. The hidden cost is not technological failure, but decision latency and talent friction created by an outdated leadership model.
Governance structures still assume that authority equals insight, and therefore the right to make the decision. AI destabilizes that equation because insight is now distributed, but authority remains concentrated. That gap is where friction accumulates.
This tension shapes how quickly organizations move, how much autonomy teams truly have, and whether leaders feel empowered or exposed. Without resolving the identity shift, no structural redesign will hold.
The coordination tension
Since late 2022, AI has been used to accelerate isolated use cases, usually for individuals or within business functions. However, enterprise value only emerges once all functions benefit together. That gap is currently where transformation stalls. Creating islands of capability within IT, finance, legal or commercial functions limits a firm’s ability to coordinate when convergence demands enterprise-wide synchronization.
Research from McKinsey, MIT, BCG and Gartner places AI transformation failure rates between 70% and 95%, attributing most failures not to technical defects but to people, culture and coordination breakdowns. In most enterprises, coordination cost is rarely measured – yet it is often one of the largest hidden expenses. Rework, duplicated analysis, delayed decisions, compliance reversals and talent churn are all symptoms of coordination failure. AI reduces processing cost but not coordination cost. Without orchestration, it almost always increases it.
When functions move faster individually but cannot integrate at the same tempo, acceleration amplifies misalignment.
The human potential tension
Efficiency gains from AI are visible, measurable and easy to celebrate. Capability erosion is invisible until it becomes expensive. When leaders automate work previously completed by people, they remove the experience that allowed sound judgment. System output might increase and the organization might move faster on paper, but beneath the surface, human confidence, trust, psychological safety and critical thinking all erode.
Organizations that lose trust and psychological safety during transformation spend far more on rework, recovery and rehiring than those that preserve human potential. Talent attrition in AI-driven transitions is frequently a coordination and communication failure, not a technological one.
Alessandro Ventura, Unilever’s VP for technology transformation, captures the challenge succinctly: “AI can be incredibly helpful doing the groundwork and providing objectivity, but the process and quality of decision-making still rest with our critical human judgment. We are a long way off AI making strategic decisions itself.”
In AI-accelerated environments, the problem intensifies because faster and more analysis does not guarantee better alignment. The harder challenge, as Massimo Muzzi, VP strategy and sustainability at ABB, observes, is “maintaining coherence at the same speed the data arrives.” AI might remove friction, but only leadership can determine whether the friction being removed was doing important work.
These three tensions – identity, coordination and human potential – are symptoms of the same underlying condition: an operating system that has not evolved as quickly as the environment it is attempting to lead.
The four conductor capabilities
In organizations that capture sustained AI value, four leadership capabilities consistently appear. They are observable patterns across leaders navigating AI transformation effectively.
1. Judgment under ambiguity
As AI makes analysis commonplace, the premium shifts from having answers to framing better questions. Leaders who excel here do not rely on speed alone. They know when to slow down, when to challenge the model, and when to override a recommendation because context, ethics or long-term consequences cannot be fully captured in data. Judgment under ambiguity is the discipline of framing better questions and clearer trade-offs.
2. Trust stewardship
In AI-enabled organizations, leaders cannot personally verify every output. Trust must therefore become systemic rather than interpersonal. Transparent reasoning, visible decision logic, and governance structures that allow safe challenge of algorithmic recommendations are structural necessities. Trust stewardship turns speed into stability.
3. Adaptive learning
Organizations that succeed with AI embed learning into the flow of work. Every decision leaves a trace. Assumptions are documented. Experiments are visible. Dissent produces refinement rather than defensiveness. Learning problems are usually system problems: fix the system and capability compounds. Organizations that treat learning as an event accumulate learning debt and build capability for a context that has already moved on.
4. Enterprise orchestration
The fourth capability is a meta-capability that maintains coherence across boundaries while ensuring context travels as fast as information
itself. It is the ability to synchronize functions without suppressing autonomy, maintain compliance without stifling experimentation, and align incentives before deployment rather than after failure.
The business case for these capabilities is clear. Organizations that cultivate them reduce coordination costs, accelerate decision velocity while preserving judgment quality, and sustain the trust that anchors performance during change. When these foundations are present, AI no longer fragments organizations through isolated optimization. It amplifies coherent enterprise performance – economically, operationally and psychologically.
Building capability in sequence
Understanding which capabilities are required is not the same as knowing how they are developed. The capabilities examined above do not appear through instruction alone. They emerge through practice, reflection and deliberate experimentation.
Across organizations navigating AI transformation, progress tends to follow a recognizable sequence. I call this developmental rhythm Spar: Strengthen, Partner, Amplify and Reshape. Spar provides a leadership pathway designed to ensure human capability evolves as quickly as the technology surrounding it – trading control for curiosity, isolated expertise for human judgment amplified by AI, and siloed thinking for collaboration and shared ownership.
Strengthen (leadership identity) Build trust through visible decision-making and transparent learning while letting go of control habits that no longer serve. When teams see your reasoning and trade-offs, they trust your judgment and learn to act with it. Strengthening identity shifts leadership from proving competence to enabling capability.
Partner (with intelligence) As AI commoditizes knowledge, value shifts from having answers to stress-testing assumptions. Treat AI as a sparring partner that exposes blind spots and expands perspective. Frame problems with better questions, ask for a challenge rather than confirmation, and integrate AI insights with human context.
Amplify (team capability) As boundaries blur, teams need distributed judgment rather than delegated tasks. When leaders model effective human-AI collaboration, teams learn integration thinking, make better cross-boundary decisions, and replicate that discipline across the organization.
Reshape (for flow) When learning and execution merge, sequential workflows become costly delays. Design work around outcomes rather than departments. The shift is from linear hand-offs to parallel collaboration, where diverse expertise converges across the value chain to create enterprise value.
The strategic implication
The prevailing narrative suggests leaders must simply understand AI better and move faster. That framing is incomplete and misleading. The challenge is architectural.
Leaders today face a choice. They can treat AI as a productivity layer added onto existing hierarchies, which will offer short-term efficiency. Or they can treat AI as a catalyst for sustained growth by redesigning how judgment, trust and coordination move through the organization. One path accelerates output. The other compounds enterprise value.
Rita McGrath, professor at Columbia Business School, puts it precisely: “Leaders who follow these principles will learn to orchestrate across that divide and, in doing so, shape not just competitive advantage, but the value and progress of organizations for future generations.”
Organizations capturing sustained AI value share a stance of curiosity over control, trust as an operating principle, and coordination as a strategic asset. They are not necessarily the ones automating most aggressively. They are the ones orchestrating most deliberately.
Without orchestration, AI’s speed fragments the organization. The question for leaders is whether they have redesigned their organizations to convert acceleration into alignment. The organizations that succeed will not be those that simply deploy more AI tools. They will be those that redesign how judgment flows, how trust is maintained and how learning compounds across their systems.
In those organizations, technology is used to elevate human capability. The result is not just faster execution, but greater coherence, stronger decision quality and a workforce that grows more capable as the technology evolves around it. The AI era does not diminish human leadership. It exposes it.
Peter Whealy is the founder of Elevate Potential and author of Lead with AI. Stay Human: How Modern Leaders Orchestrate Enterprise Value (LID Publishing)
