Why Domain Leaders Hold the Key to AI Transformation Success

AI-Capable Domain Leaders: Right Diagnosis, More Nuanced Solution Needed

The premise sounds straightforward enough: give business leaders the technical skills to drive AI transformation in their domains, and competitive advantage follows. McKinsey's recent analysis makes this case forcefully, arguing that N-2 and N-3 executives who combine traditional business acumen with AI literacy represent the scarcest and most valuable resource for digital transformation.

Their data carries weight. Among Fortune 500 senior leaders, only 17 percent of their skill sets are technical, and a mere 5 percent have held technical roles during their careers. Meanwhile, companies that successfully transform with AI consistently derive the majority of their benefits from a handful of deeply transformed domains rather than broad, shallow implementation.

The question is not whether this diagnosis is accurate. It clearly is. The more pressing question is whether the prescribed solution delivers the outcomes organizations need, and whether alternative approaches might prove more effective for different organizational contexts.

The Domain Owner Imperative: Right Problem, Complex Solution

The article's central thesis deserves serious consideration. Adam Boyd's transformation of Citizens Bank's home equity lending business illustrates the model persuasively. By developing sufficient technical literacy to oversee agile teams, understand data architecture, and drive iterative development, Boyd compressed a 35-day process to just a few days while improving customer experience and reducing costs.

This represents genuine value creation, not incremental improvement. The key insight here is that Boyd did not become a technologist. He developed enough technical fluency to ask the right questions, spot problems early, and maintain credible oversight of cross-functional teams. The distinction matters enormously.

Yet the six-to-ten hours weekly that Boyd dedicates to technology learning, multiplied across 75 to 150 domain leaders in a large organization, represents a massive investment. The opportunity cost alone demands scrutiny. Are we asking business leaders to develop a genuine second muscle, or are we asking them to become ambidextrous in ways that may compromise their core strengths?

Research from MIT Sloan Management Review on AI adoption suggests a more nuanced picture. Their 2023 study of over 3,000 organizations found that successful AI transformation requires technical literacy at multiple organizational levels, not just among domain leaders. Organizations that concentrated capability-building exclusively at senior levels achieved only marginally better results than those that distributed technical skills more broadly.

The Skills That Actually Matter

The article identifies four critical capabilities for domain owners: reimagining the domain with customers at the center, developing AI-enabled transformation roadmaps, overseeing tech delivery, and leading end-to-end change management. Each deserves examination.

The first capability, reimagining the domain, requires deep customer insight and business judgment far more than technical expertise. Menno Van der Winden's approach at Tata Steel exemplifies this: he started with the biggest operational problems, then determined whether technology could solve them. Technology followed problem identification, not the reverse.

This sequencing matters because it suggests that business acumen remains the primary skill, with technical knowledge playing a supporting role. Leaders who chase technology-first solutions consistently underperform those who pursue value-first approaches, regardless of their technical sophistication.

The second and third capabilities, developing roadmaps and overseeing delivery, present the genuine technical challenge. Here the article's prescription seems most convincing. Understanding agile methodology, data architecture, and software delivery practices enables more effective oversight and faster problem-solving.

However, alternative models exist. Harvard Business Review's research on data science project failure points to the effectiveness of dedicated translator roles—hybrid positions that bridge business and technology without requiring business leaders themselves to develop deep technical skills. DBS Bank's transformation, often cited as a success story, employed this model extensively, creating a layer of digitally-native talent that partnered with traditional business leaders rather than replacing them.

The fourth capability, leading change management, returns to traditional executive territory. The article correctly notes that domain owners have the best end-to-end view of processes being transformed, positioning them ideally to orchestrate change. This requires influence, political skill, and organizational savvy—traditional leadership attributes that technical training does little to develop.

Learning Approaches: The Hands-On Imperative

The article's most valuable contribution may be its emphasis on experiential learning over classroom training. Neesha Hathi's description of participating in design workshops, working directly with engineers and data scientists, and developing practical rather than conceptual understanding resonates with adult learning research.

The challenge is scalability. Hathi's path included running a software subsidiary and serving as chief digital officer before her current role. Few organizations can provide comparable developmental experiences to 75–150 leaders simultaneously. The two-in-a-box model mentioned briefly in the article offers one solution, pairing business leaders with technical counterparts. Yet this approach introduces its own complexities around accountability and decision rights.

The stage-gating approach advocated for managing ambiguity reflects sound practice. Breaking transformational efforts into tranches, achieving economic benefits with each phase, and maintaining flexibility to pivot addresses both the learning curve challenge and the risk management imperative. This methodology has proven effective across industries, from manufacturing to financial services.

What the article underemphasizes is the importance of psychological safety in this learning process. Senior executives accustomed to subject matter mastery often struggle with the vulnerability inherent in developing new capabilities. Organizations that create space for experimentation, normalize early failures, and celebrate learning alongside results achieve better outcomes than those that simply add technical expectations to existing performance metrics.

The Talent Dimension: Beyond Individual Capability

The article's focus on developing a nose for tech talent highlights an often-overlooked aspect of technical leadership. The performance difference between top-tier and average engineers can exceed 10x, making talent recognition and retention critical capabilities.

Yet Van der Winden's experience training over 500 engineers at Tata Steel points to something broader than individual talent assessment. He built a systematic capability development approach, using experienced team members to upskill new engineers on each use case. This suggests that organizational learning systems may matter as much as individual leader capabilities.

Walmart's e-commerce transformation offers an interesting counterpoint. Rather than exclusively upskilling existing leaders, Walmart made strategic acquisitions of digital-native companies like Jet.com, bringing in talent with both business and technical capabilities. They then facilitated knowledge transfer through integrated team structures and rotational assignments. The hybrid approach, combining organic development with strategic talent acquisition, delivered faster results than pure upskilling efforts.

The recommendation to replace 20–30 percent of current leaders deserves particular scrutiny. While clear-eyed assessment of leadership capability is essential, wholesale replacement carries significant risks. Institutional knowledge, customer relationships, and cultural understanding have value that purely technical capability cannot replace. Organizations must balance the urgency of transformation with the wisdom of institutional continuity.

Structural Enablers: The Operating Model Question

The article correctly identifies operating model changes as essential enablers of domain owner effectiveness. Embedded engineering talent, persistent funding models, and aligned performance management represent necessary conditions for success.

Yet the article treats these as relatively straightforward fixes when organizational research suggests otherwise. Shifting from project-based to persistent funding requires fundamental changes to capital allocation processes, financial planning cycles, and governance structures. Many organizations have attempted this transition and failed, not due to lack of commitment but because the ripple effects proved more disruptive than anticipated.

Similarly, embedding engineering talent under domain leader authority challenges the traditional separation between business and IT that most large organizations spent decades establishing. The benefits are real, but so are the coordination challenges, the risk of duplicated capabilities, and the potential for inconsistent technical standards across domains.

Gartner's research on fusion teams offers a complementary perspective. Rather than embedding all technical talent under domain leaders, fusion teams create integrated structures where business and technical professionals collaborate as peers with shared objectives. This model preserves some benefits of functional technical leadership while achieving the integration benefits the article advocates.

The C-Suite Role: Beyond Sponsorship

The article's recommendations for C-suite action focus primarily on enabling domain owners through talent decisions, upskilling programs, and operating model changes. While each recommendation has merit, this framing may underestimate the C-suite's direct role in transformation.

Indosat Ooredoo Hutchison's approach, described briefly in the article, hints at this broader role. CEO Vikram Sinha champions quarterly executive-suite AI immersion programs and an AI leadership initiative for the top 100 executives. This suggests that technical literacy at the C-suite level enables more effective oversight and resource allocation decisions, not just better support for domain owners.

Microsoft's transformation under Satya Nadella demonstrates this principle at scale. Nadella did not simply empower domain owners. He articulated a clear technical vision, made bold platform bets, and personally modeled technical engagement through activities like reviewing code and participating in hackathons. This top-down technical leadership complemented rather than replaced domain-level transformation.

The strategic upskilling program recommendation deserves expansion beyond the article's treatment. Effective programs require several elements:

• Clear competency models tailored to organizational context
• Multiple learning modalities addressing different learning styles
• Measurement frameworks that track application, not just completion
• Sustained commitment measured in years, not quarters

General Electric's digital industrial transformation attempt offers a cautionary tale. GE invested heavily in upskilling programs and hired thousands of data scientists and software engineers. Yet the transformation ultimately disappointed because the fundamental business strategy proved flawed. Technical capability, however well-developed, cannot compensate for strategic misjudgment.

Alternative Models and Hybrid Approaches

The domain owner model represents one pathway to AI transformation, but not the only viable approach. Product-led models, where dedicated product managers own end-to-end digital experiences, have proven effective in consumer technology companies and increasingly in traditional industries.

Amazon's working backwards methodology exemplifies this alternative. Rather than expecting business leaders to develop technical muscles, Amazon creates documents describing desired customer experiences, then assembles whatever capabilities are needed to deliver them. Technical and business expertise integrate at the team level rather than within individual leaders.

The most sophisticated organizations employ hybrid approaches that match governance models to transformation maturity. Early-stage transformations may benefit from the two-in-a-box model, pairing business leaders with technical counterparts. As technical literacy develops, organizations can transition to fuller domain owner accountability. Mature transformations may evolve to product-led models where traditional business-technology distinctions blur entirely.

Geography and industry context also matter. Highly regulated industries like banking and healthcare face constraints that pure product models struggle to accommodate. Manufacturing and industrial companies often need deeper integration of operational technology and information technology than the article's framework explicitly addresses. Global organizations must navigate varying technical talent availability across regions, affecting the feasibility of embedded engineering models.

Practical Recommendations for Implementation

For organizations convinced that developing AI-capable domain owners represents the right path, several implementation principles emerge from research and practice:

• Start with honest capability assessment. Not all current domain leaders have the aptitude, interest, or career runway to develop strong technical muscles. Forcing the model on resistant or unsuited leaders wastes resources and delays transformation.
• Create multiple learning pathways. Leaders with engineering backgrounds need different development than those from purely commercial backgrounds. Customize learning journeys based on individual gaps and learning preferences.
• Measure application, not exposure. Hours of training completed matter far less than demonstrated capability to develop roadmaps, oversee technical teams, and deliver business value through technology.
• Plan for real timelines. The six-month learning curve cited for Adam Boyd represents an optimistic scenario. Most leaders require 12–24 months of sustained hands-on experience to develop genuine fluency.
• Create communities of practice. Peer learning accelerates development and provides support networks that help leaders persist through difficult phases.
• Align incentives carefully. Leaders must see clear career benefits from developing technical capabilities; otherwise competing priorities will crowd out the sustained effort required.

Looking Ahead: The Evolving Nature of Business Leadership

The deeper question the article raises is whether we are witnessing a permanent evolution in business leadership requirements or a transitional phase until AI becomes more abstracted and easier to deploy.

History offers mixed guidance. The rise of financial management as a core leadership skill in the 1970s and 1980s proved enduring. CFO-type capabilities became expected of most senior leaders, not just finance specialists. Similarly, customer-centricity and data-driven decision-making, once specialized capabilities, are now baseline expectations.

Yet other predicted leadership evolution proved temporary or limited. The late 1990s expectation that all executives needed deep internet expertise faded as digital capabilities became more embedded in organizational structures rather than individual skill sets. The mid-2000s emphasis on social media fluency among senior leaders similarly diminished as organizations developed specialized capabilities.

AI's trajectory remains uncertain. Current complexity and rapid evolution favor the domain owner model, where business leaders develop sufficient technical literacy to make informed decisions and provide effective oversight. However, as AI platforms mature and become more accessible through natural language interfaces and no-code solutions, the technical barrier to effective use may decline.

The more enduring leadership capability may be what the article describes as reimagining domains and creating transformative vision. This requires creativity, customer insight, and willingness to challenge orthodoxies—attributes that technology cannot replicate. Leaders who combine this imaginative capability with enough technical fluency to separate realistic possibilities from hype will remain valuable regardless of how AI technology itself evolves.

Conclusion: A Necessary but Insufficient Solution

The McKinsey article identifies a genuine and important challenge: most organizations lack sufficient numbers of business leaders who can effectively bridge domain expertise with AI capability. The examples of Adam Boyd, Neesha Hathi, and Menno Van der Winden demonstrate that developing this dual capability delivers real business value.

Yet the solution requires more nuance than the article fully captures. Not all organizations need the same approach. Not all leaders can or should develop the same technical depth. And organizational structure, culture, and talent availability constrain what is feasible regardless of commitment level.

The most effective approach combines selective leadership changes with targeted capability building, supported by operating model evolution and enabled by C-suite commitment. Organizations should pilot this model in high-potential domains, learn from early experiences, and adapt approaches based on what works in their specific context.

The goal is not to create business leaders who think like technologists, but rather leaders who can work effectively with technology to solve business problems and create customer value. That distinction, while subtle, proves critical. Business judgment and customer insight remain the primary capabilities. Technical literacy serves these capabilities rather than replacing them.

For boards and C-suites assessing their transformation readiness, the article's central question deserves attention: How many domain owners with strong AI muscles does your organization have? But that question should prompt a broader discussion about organizational design, talent strategy, and transformation approach—not simply trigger upskilling programs.

The companies that succeed with AI will be those that match their leadership development investments to their strategic context, maintain realistic timelines for capability building, and create organizational structures that enable business-technology collaboration at all levels. Domain owners with strong AI muscles will be part of that formula, but only part. The harder work is building organizational systems that allow those individual capabilities to translate into sustained competitive advantage.