The Triple Return Promise Why McKinseys Enterprise Tech Vision Needs a Reality Check

McKinseys AI-Driven Tech Transformation Blueprint: A Reality Check

McKinsey has released its latest blueprint for enterprise technology transformation, arguing that companies can achieve three times the EBITDA lift from their technology investments by 2030 compared to 2025. The prescription involves six imperatives centered on harnessing artificial intelligence to reinvent technology functions. While the ambition is commendable and the strategic framework thoughtful, business leaders should approach these projections with healthy skepticism informed by decades of technology transformation promises that fell short of their forecasts.

The Core Thesis: AI Changes Everything

The McKinsey team, led by senior partners Aamer Baig, Klemens Hjartar, and colleagues, argues that artificial intelligence fundamentally alters the economics of enterprise technology. Their central claim rests on AIs ability to reduce the unit cost of introducing new functionality while simultaneously increasing engineering productivity. This creates what they describe as a virtuous cycle: improved productivity removes capacity constraints, freeing resources to modernize platforms, which further increases productivity and reduces technical debt.

The mathematical logic is elegant. Through modeling scenarios with a baseline $1 billion IT budget, McKinsey demonstrates that companies allocating 33 percent of investment to IT infrastructure in the first four years, combined with 4 percent annual budget increases, could achieve cumulative five-year EBITDA lift of $813 million. This compares to just $405 million under traditional allocation patterns that favor business-driven application development over foundational IT investment.

Yet buried within their own analysis lies a troubling admission: despite exponential improvements in generative AI coding tools, most organizations using these tools at scale have achieved less than 10 percent improvement in team productivity. This gap between theoretical potential and realized value should give pause to any executive considering a wholesale transformation based on AI productivity assumptions.

The Six Imperatives: Strategic Soundness Meets Implementation Reality

McKinseys six imperatives represent solid strategic thinking about enterprise technology management: recalibrating IT economics, rebuilding platforms, renovating enterprise data, redesigning talent models, revamping vendor relationships, and remodeling risk and resiliency. The question is not whether these imperatives make sense in theory, but whether organizations possess the capability and conditions to execute them at the scale and speed required to generate promised returns.

Consider the first imperative: recalibrating to new IT economics. The recommendation to increase overall IT budgets by 4 percent annually while shifting allocation toward infrastructure and tooling assumes several conditions that may not hold. It presumes that organizations have budget headroom at a time when many face margin pressure and cost reduction mandates. It assumes CFOs and boards will accept multi-year investment horizons before seeing material returns. Most critically, it requires that business leaders understand and support technology economics they have historically viewed with suspicion.

According to research from MIT Sloans Center for Information Systems Research, fewer than 30 percent of digital transformation initiatives achieve their stated objectives. The problem rarely stems from flawed strategy or inadequate technology. Rather, transformation failures result from organizational antibodies that reject change, misaligned incentives between business and technology functions, talent gaps, and the sheer difficulty of executing complex change while maintaining business continuity.

The Productivity Paradox Revisited

McKinseys productivity claims for AI-enabled development deserve particular scrutiny because they echo promises made during previous waves of technology transformation. During the 1990s and early 2000s, companies invested heavily in enterprise resource planning systems, customer relationship management platforms, and other enterprise software based on vendor promises of dramatic productivity improvements and cost savings. Research by economists Erik Brynjolfsson and Lorin Hitt found that while these technologies did eventually contribute to productivity gains, the returns took far longer to materialize than expected and required substantial complementary investments in organizational redesign and process change.

The productivity paradox that Brynjolfsson identified in the 1980s and 1990s refers to the phenomenon where massive technology investments failed to show up in productivity statistics for years. The eventual productivity gains came not from the technology itself but from organizational learning about how to use it effectively and from complementary innovations in business processes.

Todays AI productivity claims risk repeating this pattern. A Harvard Business School study cited by McKinsey found that developers using AI coding assistants saw their relative share of coding activities increase by approximately 5 percent while project management activities decreased by 10 percent. This suggests modest rather than transformational productivity improvements, and certainly nothing approaching the multiples required to justify the investment scenarios McKinsey models.

Furthermore, measuring software engineering productivity remains notoriously difficult. Lines of code, features shipped, and bugs fixed all provide incomplete pictures of developer value creation. The most important work developers perform often involves understanding complex business requirements, designing elegant architectures, and making trade-off decisions that affect systems for years. These activities may not accelerate meaningfully with AI assistance and may even slow down as developers must now verify AI-generated code and guard against subtle errors that automated tools introduce.

The Collaboration Conundrum

Perhaps the most valuable insight from McKinseys research comes not from their prescription but from their diagnosis of the business-technology relationship. Their interviews with over 100 technology officers revealed that only 13 percent report receiving effective support from business partners across critical interactions including long-term platform investment, risk mitigation, initiative planning, requirements definition, and ways of working.

This finding illuminates why so many technology transformations fail to deliver promised value. The problem is not primarily technical but organizational and cultural. Business leaders often treat technology as a cost center to be minimized rather than a capability to be cultivated. They demand immediate results while resisting the multi-year investment horizons that platform modernization requires. They specify vague requirements, change priorities frequently, and then blame technology teams when projects deliver late or miss the mark.

For McKinseys vision to materialize, this fundamental relationship must change. Business leaders must become genuine partners in technology strategy, willing to fund infrastructure investment that may not pay off for several years. They must invest time in defining clear requirements and accept accountability for business process changes that new technology enables. Technology leaders, in turn, must improve their ability to communicate business value in language executives understand and to deliver on commitments consistently.

Research by Jeanne Ross and her colleagues at MIT CISM suggests that companies that successfully leverage technology for competitive advantage treat enterprise architecture as a strategic asset requiring C-suite attention. These organizations maintain stable, well-funded technology platforms and resist the temptation to divert platform investment toward short-term application development. They accept that platform investment creates option value that may not be immediately visible in ROI calculations but pays dividends over time through increased flexibility and reduced time-to-market for new capabilities.

The Technical Debt Trap

McKinseys second imperative addresses technical debt, the accumulated burden of suboptimal technology decisions, deferred maintenance, and aging systems that plague most large enterprises. They claim that AI can eliminate much of the manual work in IT modernization, leading to 40 to 50 percent faster timelines and 40 percent cost reductions.

These projections warrant skepticism based on the history of automated code conversion and modernization tools. For decades, vendors have promised automated migration from mainframes to client-server architectures, from on-premises to cloud, from monoliths to microservices. While these tools provide value in specific contexts, they rarely deliver the fully automated, low-cost transformations that vendors promise.

Technical debt persists not because companies lack tools to address it but because remediation requires difficult business decisions about functionality trade-offs, acceptable risk levels, and resource allocation. AI tools may accelerate code translation, but they cannot resolve questions about which legacy systems to retire, how to migrate data with referential integrity, or how to maintain business continuity during major platform transitions.

Moreover, technical debt often reflects accumulated business logic that exists nowhere else in the organization. The mainframe code that processes insurance claims contains decades of business rules, regulatory requirements, and edge cases that may not be documented anywhere else. Translating that code without losing critical business logic requires deep domain expertise that AI tools do not possess.

Companies like Commonwealth Bank of Australia and JPMorgan Chase have spent billions of dollars and many years working to modernize core systems, even with access to cutting-edge tools and top talent. Their experiences suggest that platform modernization remains expensive, risky, and time-consuming regardless of available technology.

The Talent Model Transformation

McKinseys fourth imperative calls for redesigning talent models around human-agent collaboration. This vision of engineers building and commanding fleets of AI agents represents an interesting future state, but it assumes capabilities and organizational structures that few companies currently possess.

The concept of artisanal engineers teaching agents the craft of engineering reveals both promise and peril. On one hand, capturing expert knowledge in systems that can scale represents genuine value creation. On the other hand, this approach assumes that engineering expertise can be codified and transferred to AI systems more easily than evidence suggests.

Software engineering involves tacit knowledge that experts struggle to articulate explicitly. Experienced engineers make thousands of micro-decisions based on pattern recognition, intuition, and contextual understanding that may not translate readily into agent training. The apprenticeship model McKinsey proposes could take years to yield results and may produce agents that amplify the biases and limitations of their human teachers.

Furthermore, the talent implications of widespread AI adoption in software development remain uncertain. If AI dramatically increases individual productivity, do companies need fewer engineers or do they expand ambitions to match expanded capacity? History suggests the latter. Previous waves of productivity-enhancing tools in software development led not to workforce reductions but to more ambitious projects and expanding software scope. Jevons paradox, which McKinsey references, suggests that improved efficiency increases total consumption rather than reducing it.

Vendor Dynamics and Lock-In Risk

The fifth imperative addresses vendor relationships, noting that AI is changing dynamics between buyers and sellers across semantic, workflow, AI platform, infrastructure, and service delivery layers. McKinsey suggests that companies can replace SaaS vendors with agentic workflows and secure more favorable terms from IT service providers.

This analysis underestimates the sophisticated strategies that vendors employ to create and maintain lock-in. Major cloud providers, SaaS vendors, and enterprise software companies have spent years developing ecosystem advantages that extend far beyond individual products. They offer integrated suites where components work together seamlessly, making it costly to mix and match. They provide training, certification, and partner networks that create organizational dependencies. They continuously innovate to stay ahead of alternatives.

The idea that companies can replace established SaaS vendors with homegrown agentic workflows assumes several questionable propositions. First, it presumes that building and maintaining internal solutions costs less than subscribing to proven products. Second, it assumes that companies possess sufficient engineering talent to build these solutions while also executing on their core product roadmaps. Third, it ignores the opportunity cost of engineering time spent on undifferentiated internal tools rather than competitive capabilities.

Research by Bain & Company on cloud economics found that many organizations that built private clouds or repatriated workloads from public clouds discovered that total cost of ownership exceeded expectations once they accounted for talent costs, opportunity costs, and the pace of innovation at major cloud providers. The same dynamics may apply to replacing SaaS with internal agentic workflows.

The Risk and Resiliency Imperative

McKinseys final imperative addresses risk and resiliency, noting that AI both enables new security capabilities and introduces new vulnerabilities. This represents perhaps the most understated challenge in their framework.

As companies increase dependence on AI systems for core functions, they face several risk categories that remain poorly understood. AI systems can produce plausible but incorrect results with confidence, making errors harder to detect than traditional software bugs. They can be manipulated through adversarial inputs in ways that conventional security approaches struggle to prevent. They introduce supply chain risks through dependencies on large language models and other foundation models controlled by a small number of providers. They create bias and fairness concerns that carry regulatory and reputational risks.

The suggestion that companies can build safe AI deployments by reinventing threat modeling for agentic development and training agents to align with organizational values makes this sound simpler than reality suggests. AI alignment represents one of the fundamental unsolved problems in computer science. Major AI research labs struggle to ensure that their models behave safely and align with human values. Expecting individual enterprises to solve these problems for their internal agentic systems seems optimistic at best.

Furthermore, the regulatory environment around AI continues to evolve rapidly. The European Unions AI Act, emerging US regulations, and sector-specific requirements create compliance obligations that companies must navigate. Building extensive agentic systems now risks creating technical debt in the form of systems that violate future regulations or require extensive retrofitting to achieve compliance.

The Implementation Gap

The gulf between McKinseys strategic vision and organizational reality emerges most clearly in their own findings about current practices. Despite near-universal agreement among technology leaders that these imperatives are critical, few apply them at scale. This gap between knowing what to do and doing it represents the central challenge of enterprise technology management.

Implementation failures stem from multiple sources. Organizations lack the talent to execute sophisticated technology strategies at scale. A survey by Harvey Nash found that 65 percent of technology leaders report skills shortages in critical areas including AI, cybersecurity, and cloud architecture. These shortages constrain execution regardless of strategy quality.

Organizations also struggle with the sustained focus and investment discipline that platform modernization requires. Quarterly earnings pressures and leadership turnover create incentives for short-term thinking. New executives want to make their mark quickly, leading to strategy shifts that disrupt multi-year technology initiatives.

The organizational change management required to capture value from new technology capabilities often receives insufficient attention and investment. Technology implementations that ignore the human and process dimensions of change typically fail to deliver expected benefits. Research by John Kotter found that fewer than 30 percent of change initiatives succeed, with most failures resulting from organizational rather than technical factors.

A More Realistic Path Forward

Despite these challenges, McKinseys article contains genuine strategic insights that business leaders should consider, albeit with expectations tempered by implementation realities.

First, the core insight about shifting investment toward technology platforms and infrastructure deserves serious consideration. Organizations that underfund platforms create technical debt that constrains future capability development. However, realistic expectations about timeframes and returns are essential. Platform investments typically require three to five years before delivering material benefits, and expected returns should be measured in increments of 20 to 40 percent rather than multiples of two or three.

Second, the emphasis on business-technology collaboration identifies a critical gap in most organizations. Closing this gap requires structural changes including technology representation in business strategy development, joint accountability for business outcomes, and incentive alignment. Companies like Capital One and USAA have demonstrated that treating technology as a core business capability rather than a support function can create competitive advantage, but the transformation requires years of sustained effort and leadership commitment.

Third, the focus on data quality and AI readiness addresses a genuine organizational need. However, data remediation remains labor-intensive and politically challenging. Different business units often define metrics differently, have conflicting data quality standards, and resist standardization that might reveal uncomfortable truths about their operations. Progress requires executive sponsorship and sustained investment, with realistic expectations that comprehensive data remediation takes years rather than quarters.

Fourth, the call to revamp vendor relationships reflects important shifts in technology economics. Companies should actively manage vendor relationships, understand lock-in risks, and negotiate from positions of informed strength. However, this differs from the more aggressive posture of replacing established vendors with internal solutions. A more pragmatic approach involves using open standards where possible, maintaining multi-vendor strategies in critical areas, and building internal capabilities that increase negotiating leverage without attempting to replicate all vendor functionality.

Conclusion: Promise Meets Pragmatism

McKinseys vision of tripling enterprise technology ROI through AI transformation presents an aspirational goal that can help organizations think more strategically about technology investment. However, business leaders should approach these projections as directional rather than predictive.

The history of enterprise technology contains numerous examples of promised transformations that delivered partial results over longer timeframes than forecast. Enterprise resource planning, service-oriented architecture, cloud computing, and other technology waves each promised revolutionary benefits. Some delivered substantial value, but typically over decades rather than years and at costs exceeding initial projections.

AI represents a genuinely important technological shift with potential to increase productivity and enable new capabilities. Organizations that invest thoughtfully in AI capabilities while maintaining realistic expectations about timeframes and returns will likely outperform peers who ignore these trends. However, the gap between theoretical potential and realized value depends heavily on execution capabilities, organizational readiness, and sustained investment discipline.

Business leaders should focus on several practical priorities rather than betting entirely on McKinseys ambitious projections:

• Incrementally increase platform investment while measuring impact rigorously.
• Start with contained experiments that can demonstrate value and build organizational confidence.
• Invest seriously in business-technology collaboration through structural changes, not just exhortations.
• Develop talent strategies that combine hiring, development, and thoughtful use of AI tools to augment rather than replace human judgment.
• Manage vendor relationships actively while avoiding the trap of trying to build everything internally.

Most importantly, recognize that technology transformation remains fundamentally a leadership challenge rather than a technical one. The companies that extract superior value from technology investments do so not because they have better tools or strategies but because they build organizational cultures that view technology as strategic, maintain investment discipline over multiple years, and create genuine partnership between business and technology leadership.

The promise of tripling technology ROI should inspire ambition while humility about implementation challenges ensures that ambition translates into realistic plans. Organizations that balance these perspectivesthinking boldly about potential while planning pragmatically for executionposition themselves to capture meaningful value from enterprise technology investment in the AI era. Those that bet entirely on consulting models without accounting for organizational realities risk joining the long list of transformation initiatives that promised revolution but delivered disappointment.