For successful AI adoption, leaders need to focus on a consideration all too often viewed as secondary
Writing Edosa Odaro
p>Artificial intelligence now sits at the center of many corporate strategies. Boards expect it to unlock productivity, improve customer experience and create new sources of growth. Organizations are investing heavily in data platforms, machine learning models and generative AI tools in the hope of gaining competitive advantage. Yet despite impressive technological progress, many AI initiatives struggle to produce the value leaders expect. McKinsey’s The State of AI report consistently finds that while AI adoption has accelerated across industries, only a relatively small share of companies report capturing substantial financial value from their AI deployments. The algorithms work. The models perform. Tech capabilities improve. But somewhere between technological success and organizational impact, the expected value often fails to materialize.
Studies from MIT Sloan Management Review and Boston Consulting Group point to a similar challenge: while many organizations experiment with AI, only a small proportion succeed in scaling it to deliver meaningful business impact. Leaders find themselves asking a persistent question: why does so much promise produce so little visible value? The answer may lie not in the technology itself but in the way organizations define value.
The hidden complexity of AI value
When leaders discuss artificial intelligence, the conversation typically begins with purpose: Why are we investing in AI? Why does this initiative matter now? Why will it improve our competitive position?
This way of thinking reflects a widely adopted leadership principle – begin with purpose before defining action. In many situations, this works well. A clear sense of purpose can align teams, inspire commitment and guide strategic choices. AI, however, introduces a complication that traditional leadership models rarely address.
AI systems rarely affect only one group. Unlike many traditional technology investments, AI reshapes outcomes across multiple stakeholders simultaneously. A single AI initiative may affect employees, customers, shareholders, regulators and society at large. When this happens, value is not simply created. It is redistributed.
Consider a few familiar examples. A customer service chatbot may allow an organization to respond to queries more quickly, creating value through speed and efficiency – but at the same time, reducing the role of human agents, reshaping employment and changing the nature of customer relationships. An automated pricing system may optimize revenue and improve margins for shareholders – yet customers may experience the resulting price adjustments very differently. A predictive hiring system may allow organizations to screen candidates faster and identify useful patterns in recruitment data – but the criteria embedded within that system may also influence who gains access to opportunity.
In each case, the technology may perform exactly as intended, but the value created by the system will be experienced differently depending on who is affected. In that sense, AI does not only automate decisions. It also automates the priorities and assumptions embedded within those decisions.
Why the value question becomes difficult
This multi-stakeholder reality creates a challenge that many organizations underestimate. Different groups often define value in very different ways. For executives, value may be measured through growth, productivity or market advantage. For employees, it may be linked to meaningful work, professional development or job security. For customers, it may involve fairness, trust and experience. For regulators and society, it may encompass accountability, transparency and safety. When AI systems are introduced, these perspectives can intersect in complex ways; an initiative that appears successful from one vantage point may create tensions from another.
Yet in practice, many organizations proceed as if value were a single, universally shared objective. AI strategies are often designed around capabilities, roadmaps and technical milestones without fully exploring how the resulting value will be distributed among stakeholders. This can create a subtle but powerful misalignment. Technical teams focus on model performance and scalability; business teams focus on operational outcomes; executive leadership focuses on strategic positioning. Meanwhile, the broader implications of the system remain implicit.
The result is not necessarily failure in the technical sense: the models may perform well and the infrastructure may operate effectively. But the organization struggles to convert those capabilities into value that is clearly recognized, trusted and supported across stakeholders.
The question most AI strategies avoid
At the heart of this challenge lies a question that many AI strategies overlook. When artificial intelligence creates value, who benefits from that value? And equally important: who might bear the costs?
These questions are not always comfortable, because they reveal the trade-offs embedded in technological progress. AI systems often improve efficiency by automating tasks previously performed by people, increase predictive accuracy by analyzing patterns in historical data, and optimize decisions by prioritizing certain outcomes over others. Each of these design choices reflects a set of priorities. AI does not simply automate decisions: it scales the priorities embedded within them, and those priorities inevitably influence how value is distributed.
The important point is not that these trade-offs exist; trade-offs are inherent in every strategic decision. The problem arises when organizations fail to examine them consciously. When assumptions about value remain implicit, they become embedded within systems by default rather than by design. Once operationalized through algorithms and automated processes, these assumptions can shape outcomes at scale. By the time tensions become visible – through employee resistance, regulatory scrutiny or public concern – the system may already be deeply embedded in operations.
Why organizations avoid the question
There are understandable reasons why organizations hesitate to confront the distribution of value in their AI initiatives. The first is speed. AI has moved rapidly from experimental technology to strategic priority. The pace of adoption has accelerated dramatically; the Stanford AI Index reports sustained year-on-year growth in global AI investment and enterprise deployment across nearly every industry. Boards expect visible progress, competitors announce ambitious AI programs, and leaders feel pressure to demonstrate momentum. In such an environment, implementation often takes precedence over reflection.
The second reason lies in organizational structure. AI initiatives frequently originate within technical or data teams who naturally focus on model performance, data quality and system scalability. While these are essential considerations, they do not always capture the broader organizational consequences of the systems being developed.
The third reason is accountability. Questions about who benefits and who might lose from AI often cross organizational boundaries, and no single department fully owns these outcomes. As a result, the broader implications of AI systems can remain unexamined.
This is not necessarily the result of poor leadership – in many cases, it reflects the complexity of the systems themselves. Artificial intelligence introduces subtle shifts in how decisions are made, how work is organized and how value is distributed. But avoiding these questions carries its own risks. When trade-offs are not examined explicitly, they tend to surface later, when the costs of putting things right may be much higher. Addressing these questions early on allows leaders to shape AI initiatives deliberately rather than reacting to consequences after they appear.
A different starting point
When it comes to AI, leaders should not start with purpose – asking why a system should exist – but with an understanding of who it would affect. The priorities: who, what, and then why.
- Start with who Begin with stakeholders. Who benefits from the value the system creates? Who might bear the cost of that value? Whose interests must be safeguarded? Understanding these stakeholder dynamics reveals the broader context in which the system will operate and makes it possible to define the system’s purpose in a meaningful way, rather than in the abstract.
- Define what Once the stakeholder landscape is clear, the ‘what’ will be clearer. What capability should the AI system deliver? What decisions will it influence? What outcomes should it improve? These questions can now be answered with the stakeholder context in mind, rather than in isolation from it.
- Clarify why Only after these questions are addressed does the final question, the why, emerge. Why does this initiative matter strategically? Why is it important for the organization’s long-term direction?
This sequence does not diminish the importance of purpose. Instead, it recognizes that in the context of AI, purpose cannot be defined meaningfully without first considering the stakeholders involved.
Leadership responsibility
Because AI intelligence redistributes value across stakeholders, AI strategy is fundamentally a leadership responsibility. Technology teams can build powerful systems, data scientists can develop accurate models and engineers can scale infrastructure – but determining how these capabilities translate into value requires leadership judgment. Three responsibilities are particularly important.
Value alignment Leaders must ensure that AI initiatives connect clearly to the outcomes the organization seeks to achieve, translating technical capabilities into language that resonates across business units, operational teams and executive leadership.
Trade-off visibility Every AI system reflects priorities embedded in its design. Leaders must make those priorities explicit: who benefits from the system, who might experience unintended consequences, what safeguards are necessary to maintain trust. Making these considerations visible does not slow innovation; it strengthens the foundation on which innovation occurs.
Cross-stakeholder governance Because AI initiatives influence multiple groups simultaneously, they require oversight that extends beyond any single department. Effective governance brings together technical expertise, business leadership and broader stakeholder perspectives.
When these three elements are aligned, AI becomes more than a technological capability – it becomes a coordinated effort to create value that is both economically meaningful and socially sustainable.
From technology strategy to value strategy
Artificial intelligence will undoubtedly reshape many industries in the years ahead. But the organizations that succeed will not necessarily be those with the most advanced models or the largest data platforms.
They will be the ones that recognize a simple truth: AI does not only create value. It redistributes it. Leaders who acknowledge this reality early can design systems that align value across stakeholders rather than allowing unintended consequences to emerge later.
In doing so, they move the conversation beyond technology. AI becomes not just a tool for automation, but a strategic instrument for shaping how value is created, shared and sustained. And that process begins with a question that many AI strategies still avoid: not simply why an AI system should exist, but who it is truly designed to serve.
Edosa Odaro is an AI and data transformation leader and author of The Values of Artificial Intelligence: How Smart Leaders Capture and Connect AI Value to Human Values (Auerbach Publications)
