Technology Doesn’t Scale, Culture Does

By Dominique Shelton Leipzig, Founder & CEO of Global Data Innovation; James D. White, Executive Mentor, Leader, Cultural Transformation Practice, ExCo Group, and Chair, Board of Directors, the Honest Company; Keith Meyer, Senior Managing Director and Co-Lead of the Board Impact Practice, ExCo Group; and with contributions from Natalie Rothman, Chief Human Resources Officer at The Hershey Company

Many organizations expected artificial intelligence (AI) to unlock rapid productivity gains, enhance decision-making, and create new sources of value. Yet organizations continue to face a persistent challenge: while investment in AI has accelerated, measurable enterprise-wide returns have seemed to lag.

The primary barrier is rarely technological. More often, it is one of culture and trust. Many organizations have not fully aligned leadership, governance, workforce capabilities, and operating models to support AI at scale. Without a shared understanding of how AI should be used, governed, and integrated into decision-making, organizations may struggle to move beyond isolated pilots and achieve sustained impact.

AI is no longer a discrete technology initiative; it is a strategic and organizational transformation that depends on trust across leadership teams, within the workforce, and with external stakeholders. Boards are increasingly expected to move beyond oversight and serve as active partners in shaping long-term value, ensuring that organizations build cultures capable of adopting AI responsibly and effectively.

At the same time, governance expectations continue to expand. Directors should help ensure that organizations establish trustworthy foundations for AI, grounded in high-quality data, clear accountability, effective oversight, and sound human judgment. Trust is not a byproduct of AI adoption; it is a prerequisite. Without it, confidence can erode quickly, slowing adoption and increasing risk.

Organizations most likely to succeed with AI are those that design cultures that equip employees with the right level of understanding, align leaders around thoughtful governance, and build the capabilities required to experiment, learn, and scale AI responsibly.

As Natalie Rothman, Chief Human Resources Officer at The Hershey Company, shared, “The most consequential decisions we’ve made have not been about which models to deploy. They’ve been about how we bring our people along, how we define accountability, and how we measure whether this is creating value.”

The Question Boards and CEOs Are Asking About AI

Today, a consistent set of questions continues to surface in boardrooms:

• Why has AI not yet delivered the level of return initially expected?
• Why do so many AI initiatives remain in pilot mode?
• How can my organization scale AI beyond isolated use cases?
• Do we have the right leadership and capabilities to govern AI?
• Are we effectively managing the cyber, regulatory, and reputational risks associated with AI?

These questions reflect a broader reality: AI is not solely a technical challenge. It is a leadership, governance, and operating model challenge—one defined by the ability of the organization to operate in an environment of continuous innovation and disruption.

The AI Value Gap

A common pattern has emerged across organizations: AI initiatives often begin within technical teams, while critical stakeholders across finance, operations, legal, human resources, and the board become involved later in the process. This sequence is not simply a workflow issue—it reflects a deeper cultural dynamic, where AI is treated as a specialized initiative rather than an enterprise-wide capability.

In practice, successful organizations tend to engage cross-functional leaders from the outset. CFOs help determine which AI use cases align with strategy and value creation; HR leaders shape workforce education and reskilling; operational experts define real-world constraints and accuracy thresholds; and legal and compliance leaders establish guardrails for experimentation and risk. When this alignment is absent, organizations may lack a shared understanding of how AI should function within the enterprise.

The consequences are predictable. Companies often evaluate return on investment after initiatives are already underway, rather than design use cases around strategic priorities, financial objectives, and operational realities from the beginning. Metrics tend to focus on activity—productivity gains, pilots launched, tools deployed, models tested—rather than business outcomes that build trust and credibility.

Closing this gap requires redefining what value looks like. Organizations should move from isolated productivity gains to enterprise-level value creation. Automating workflows, not just activities, becomes essential. For example, at Hershey, Rothman shared that early AI use cases focused on summarizing meetings or drafting emails. But more consequential gains emerged when AI was embedded into operational workflows, such as demand planning, content creation at scale, commercial investment modeling, and retail sales intelligence. Each use case tied directly to decisions the organization already owns and measures.

Hershey’s broader approach reflects this progression. It operates on three levels simultaneously: employee digital fluency as the baseline, functional workflow automation as the mid-layer, and cross-functional enterprise priorities where AI creates differentiated advantage. As Rothman shared, “The vision is simple but deliberate: make AI ordinary and powerful—not aspirational, but embedded in how work gets done every day.”

Importantly, this progression is reinforced through operating model design. When HR leaders are intentionally embedded alongside technology leaders within each function—and when digital advocates are driving adoption on the ground—a continuous, real-time feedback loop around AI is created. This makes it possible to pinpoint where friction is emerging and to distinguish whether it stems from cultural factors, such as mindsets and ways of working, or from capability gaps that require targeted skill-building.

Rothman also noted that at Hershey, this approach began with a deliberate decision to assess before accelerating—establishing governance structures, redesigning risk exposure processes, and treating upskilling as a foundational element of the strategy, not an afterthought.

For boards and executives, this creates a clearer lens on readiness. Leaders can assess whether their organization is culturally prepared for AI by asking the questions below. If these questions trigger silence in the boardroom, the issue is rarely technology—it is leadership alignment and culture.

1. Does our CFO help determine which AI use cases are pursued before projects begin, or only evaluate them after they are built?
2. Do operational experts provide accuracy benchmarks, workflow realities, and real-world constraints to AI development teams?
3. Can our board clearly explain how AI systems are monitored after deployment and how incidents escalate?
4. Do cross-functional AI governance meetings produce decisions or simply discussions?
5. Have we established a baseline level of AI education for employees, with role-specific trainings for managers, technical teams, and high-risk functions?
6. Do employees know where experimentation is encouraged, where human review is mandatory, and how to flag concerns when AI outputs are incorrect?
7. Is there a plan to train the entire board on AI oversight to enhance innovation and minimize risk?
8. Does the board have an effective framework to evaluate management’s responses to the questions above?

Building Trustworthy Foundations for AI

As AI adoption scales, trust becomes central. Boards are increasingly focused on whether organizations have established the foundations necessary to deploy AI responsibly and effectively.

This begins with data. Quality, accuracy, and governance of data are fundamental to AI outcomes. Weak data foundations may introduce risk not only to performance, but to compliance, reputation and stakeholder trust.

Equally important is ongoing monitoring. AI systems require continuous evaluation to identify bias, accuracy issues, cybersecurity risk, and performance drift over time. Yet error rates remain significant, and they are not theoretical—they are already impacting real-world operations.

The result is not only a gap between investment and value realization, but also a broader “say/do” gap. Leaders emphasize the importance of trustworthy AI, yet organizations often lack the routines, capabilities, and accountability needed to deliver it. This reflects a cultural disconnect between ambition and execution, and between governance principles and day-to-day behaviors.

To build a trustworthy foundation, accountability is critical. Rather than relying on isolated roles, organizations are moving toward integrated models in which responsibility for AI is shared across technology, business, legal, and risk functions. Cross-functional mechanisms, such as AI councils, can support coordination and decision-making, but they require clear objectives and operating frameworks. Otherwise, they risk eroding trust by consuming time without driving meaningful outcomes.

For boards, the lesson is clear: AI deployment is not simply a technology challenge—it is a governance, capability-building, and leadership challenge.

Forward-looking boards are moving beyond activity reports toward outcome-based oversight, and the ones doing it well are asking fundamentally different questions. Without a structured scorecard, there is no transparency. Without transparency, there is no real accountability. Boards should be asking—and management should be reporting—whether AI training has changed how work gets done, and whether those changes have produced measurable outcomes. Transparency through the scorecard is what makes those questions answerable.

AI is Reshaping C-Suite Structures and Organizational Design

The governance imperative is to keep humans meaningfully integrated in the loop—not as a compliance requirement, but as a design principle, according to Rothman.

Traditional functional boundaries are beginning to blur, particularly across technology leadership roles. CIO, CTO, and AI leadership responsibilities are increasingly integrated, reflecting the need for coordinated execution. Hybrid roles, like Chief AI Strategy Officer or Trust Officer, are emerging, combining technical expertise with strategic and operational responsibility.

At the same time, the role of the CHRO is expanding to include not only talent management, but also workforce transformation, culture, and organizational design. The board’s role in workforce transformation is evolving, as well. While many boards can tell you how much their organization has spent on AI, fewer can clearly articulate how the workforce must be structured to use AI effectively.

The future workforce will increasingly manage AI agents, directing systems that execute tasks, surface insights, and flag exceptions. New roles will sit across business functions, blending deep domain expertise with human judgment. This represents a fundamentally different operating model, and organizations that design it intentionally will more likely outpace competitors.

Successful AI deployment requires collaboration across multiple disciplines:

• Technical teams, who build the systems
• Operational experts, who understand real-world workflows
• Legal and compliance leaders, who manage regulatory exposure
• Financial executives, who measure value creation
• HR and learning leaders, who drive AI literacy, reskilling, and adoption support

As AI accelerates toward flatter, more adaptive organizational structures, collaboration, education, and reinforcement across functions become essential to execution.

Principles to Practice: The T.R.U.S.T. Approach

It is important for leaders to first understand how AI is changing the work of frontline employees, middle managers, customers, and experts across the business. Leaders should then identify the most important capability gaps, governance risks, and adoption barriers to solve.

Organizations that successfully deploy AI share a common set of leadership behaviors: T.R.U.S.T.

• Triage: Identify AI use cases aligned with strategic priorities, risk tolerance, and capabilities the organization is prepared to support.
• Right Data: Ensure training data is accurate, ethical, and legally appropriate, while drawing in the operational expertise that gives the data business meaning.
• Uninterrupted Monitoring: Continuously test AI systems for accuracy, biases, cybersecurity vulnerabilities, operational risks, and adoption breakdowns after deployment.
• Supervision: Maintain human oversight, capable of identifying and correcting deviations, and train employees on when to rely on AI, verify it, and stop it.
• Technical Documentation: Develop documentation that allows AI systems to be diagnosed, corrected, improved, or shut down when necessary, while preserving institutional learning over time.

When implemented effectively, the T.R.U.S.T. approach becomes a shared language across the organization—connecting technology teams, operational leaders, legal counsel, HR, financial executives, managers, and boards. This alignment enables organizations to move from fragmented experimentation to disciplined enterprise deployment. It also helps improve the organization’s capabilities by making teams better at problem-solving, faster at learning, and clearer about accountability.

From Audit to Capability Building

Achieving AI success requires leadership to adopt an “AI success blueprint.” To start, organizations should conduct a culture audit to determine whether cross-disciplinary collaboration exists or whether AI governance simply reinforces existing silos. Next, leaders should assess AI literacy across the enterprise. Not every employee needs to become an engineer, but all need a working understanding of what AI is and is not, where it can help or harm, and what successful and trustworthy use looks like in their roles.

Thoughtful governance does not shut down experimentation or slow down innovation. It creates the conditions to accelerate innovation by facilitating safe learning, useful piloting, and disciplined scale. Therefore, leadership teams should redesign AI governance to ensure that training, experimentation, and escalation are built into the operating model. By extension, middle managers must be equipped to bring AI adoption to life. They are the hinge between strategy and execution, translating principles into routines, coaching employees through change, and surfacing what is happening on the ground.

Finally, organizations need ongoing training and feedback loops. AI adoption is iterative. New models, regulations, and workflows require repeated communication, refreshed learning, and visible reinforcement from senior leadership.

Conclusion

The challenge facing organizations today is not adopting AI, but integrating it effectively into the enterprise. The organizations that lead the AI era will not necessarily be those with the most advanced algorithms, but those whose leaders build cultures that enable trust and scale—educating their people, aligning leadership around disciplined and transparent governance, and developing the capabilities to experiment, learn, and adapt continuously. In these organizations, culture becomes the primary enabler of value, and leadership the multiplier: CHROs, CTOs, CFOs, commercial leaders, and general counsels operate as true transformation partners, embedding AI into how work gets done and creating the conditions for sustained enterprise value.

To receive exclusive corporate governance insights for board members and leaders, join the Nasdaq Center for Board Excellence.

The views and opinions expressed herein are the views and opinions of the authors and do not necessarily reflect those of Nasdaq, Inc.