Why Strategic Focus Not Broad AI Deployment Separates Winners from Losers

By Staff Writer | Published: June 17, 2026 | Category: Strategy

As AI investment accelerates across industries, the most dangerous mistake executives can make is treating it as a technology problem rather than a strategic and organizational one.

Focused AI Beats AI Everywhere

There is a particular kind of organizational hubris that emerges during periods of technological excitement. Boards pressure CEOs to move fast. Consultants arrive with roadmaps. Vendors promise seamless integration. And somewhere in the middle of all this momentum, companies lose the one thing that separates transformation from theater: strategic focus.

Tanguy Catlin, McKinsey senior partner and contributor to the second edition of Rewired, makes a case that is at once commonsense and routinely ignored. Drawing on insights from more than 200 enterprise-wide tech and AI transformations globally, Catlin argues that the organizations achieving genuine, lasting results are not the ones deploying AI across every function simultaneously. They are the ones that begin with a clear-eyed decision about where AI will create the greatest and most distinctive value, then build deeply in those one to three domains before expanding further.

This is not a novel argument in the abstract. Michael Porter wrote about the dangers of competing on all fronts decades ago. But in the context of AI adoption—where the pressure to demonstrate broad capability is relentless and where the technology itself seems endlessly applicable—the discipline required to stay focused is genuinely difficult. Catlin's contribution is in grounding that discipline with data and in connecting it to an equally important insight: technology alone does not create transformation. Process redesign must come first.

The Argument for Focused AI Investment

Catlin's core thesis is that AI transformation requires organizations to make deliberate choices about where they want to be truly distinctive. The data supporting this position is compelling. McKinsey's research across hundreds of transformations consistently shows that organizations attempting enterprise-wide AI deployments from the outset struggle to capture measurable value, while those that concentrate resources on a small number of high-priority domains generate disproportionate returns.

This mirrors patterns observed in other technological transitions. When cloud computing emerged as a strategic imperative, the companies that captured early competitive advantages were not those that migrated everything to the cloud simultaneously, but those that identified the specific workloads and customer-facing capabilities where cloud infrastructure would change the competitive equation, and then executed with precision.

The same logic applies to AI, but with an important added dimension. Because AI is a general-purpose technology, the temptation to use it everywhere is even stronger. A large financial services company, for instance, might identify genuine AI use cases in underwriting, claims processing, customer service, fraud detection, regulatory compliance, and portfolio management, all at once. The technology could theoretically improve all of these areas. But attempting to transform all of them simultaneously fragments leadership attention, stretches talent pools thin, creates competing priorities in change management, and ultimately produces marginal improvements across the board rather than transformative impact in the areas that most determine competitive position.

This perspective aligns with research published in MIT Sloan Management Review, which found that firms achieving the highest returns from AI investments were significantly more likely to have a defined AI strategy tied to specific business outcomes rather than a broad portfolio of disconnected initiatives (Ransbotham et al., 2020). The study, which surveyed more than 3,000 managers across industries, noted that strategic clarity about AI priorities was a stronger predictor of value capture than the scale of AI investment itself.

Process Redesign Before Technology Enablement

Perhaps the most practically important element of Catlin's argument is the sequencing he insists upon: redesign processes first, then enable them through technology. This sounds obvious. In practice, it is almost universally reversed.

Organizations typically acquire AI tools or platforms, then ask how existing processes can be automated or augmented using those tools. The result is what might be called digital veneer: a layer of technology applied over processes that were designed for a pre-AI world, producing incremental efficiency gains rather than structural transformation.

Consider the insurance industry, Catlin's primary domain of expertise. Many insurers have deployed AI-driven document processing tools to accelerate claims handling. The technology is genuinely capable of extracting and categorizing information from complex documents far faster than human processors. But if the underlying claims process still requires the same handoffs, approval chains, and exception-handling logic that it did before, the AI tool simply makes the existing process run slightly faster. It does not change what is possible.

Contrast this with an insurer that begins by asking a fundamentally different question: if we could design the ideal claims experience from scratch, knowing what AI can now do, what would that process look like? The answer might involve real-time data gathering at the point of incident, automated triage and settlement for straightforward claims, dynamic escalation protocols for complex cases, and a dramatically reduced human workforce concentrated entirely on the exceptions that genuinely require judgment. That is a transformed process. The AI is then built to enable it, rather than applied as an afterthought.

This sequencing principle has been validated by research from MIT's Center for Information Systems Research, which found that companies achieving what researchers termed IT-enabled organizational transformation consistently started with business process redesign rather than technology deployment, and treated the technology as an enabler of a new operating model rather than an improvement to the existing one (Ross, Weill, and Robertson, 2006; updated analysis 2024).

The Three Lenses Every CEO Must Hold Simultaneously

One of the most instructive elements of Catlin's perspective is his description of the challenge facing CEOs attempting to navigate AI transformation. He identifies three simultaneous lenses that leaders must manage: a strategic lens examining how AI will reshape industries and competitive dynamics, a technology lens governing investment and build-versus-buy decisions, and a people lens addressing how the workforce will adopt new ways of working.

The difficulty, as Catlin acknowledges, is that all three are moving at the same time. Most organizational structures are not designed to integrate these lenses coherently. Strategy teams think about competitive positioning. IT departments make technology investment decisions. HR organizations manage workforce planning and change. In too many companies, these functions operate in parallel rather than in genuine alignment, producing AI initiatives that are strategically coherent but technically flawed, or technically excellent but organizationally stranded.

The organizations that are succeeding at AI transformation tend to have broken down these functional silos at the leadership level. They have created governance structures—whether formal AI councils, dedicated transformation offices, or redesigned executive committee accountabilities—that force the integration of strategic, technical, and organizational thinking before major decisions are made.

JPMorgan Chase offers a useful reference point. The bank's AI transformation journey, which has included the deployment of AI across trading, risk, fraud detection, and customer service functions, has been notable not just for its scale but for the way it has been governed. The bank created a dedicated AI and machine learning center of excellence that explicitly integrates data science, technology infrastructure, business unit strategy, and change management, treating these not as sequential handoffs but as concurrent workstreams that inform each other continuously (Davenport and Mittal, 2022).

The Hardest Part: Making Transformation Stick

Catlin reserves his most pointed observation for last. Strategy matters. Technology matters. But change management, he argues, is the hardest part. The real challenge lies not in identifying the right AI applications or building the right models, but in genuinely aligning the organization and ensuring that transformation endures beyond the initial deployment.

This is where most AI transformations fail, and the evidence is substantial. A study by Boston Consulting Group found that approximately 70% of digital transformation initiatives fail to achieve their stated objectives, and that the primary causes of failure are organizational rather than technical: lack of clear ownership, insufficient leadership alignment, inadequate capability building among frontline employees, and a failure to embed new ways of working into the culture and incentive structures of the organization (BCG Henderson Institute, 2020).

The pattern is familiar to anyone who has observed large-scale organizational change programs over the past two decades. A company invests significantly in a new technology platform, achieves a technically successful implementation, and then watches adoption stagnate because the people who are supposed to work differently have not been adequately prepared, motivated, or supported to do so. The technology sits underutilized while the organization reverts to familiar behaviors.

Catlin's insistence on change management as the central challenge is both correct and, in certain respects, incomplete. Change management as a discipline has often been conceptualized as a communication and training challenge: explain the change, train people on the new tools, manage resistance. But AI transformation requires something more fundamental: a shift in how employees understand their own roles and value within the organization.

When AI automates significant portions of what a workforce previously did manually, the psychological contract between employer and employee changes. People need to understand not just how to use new tools but why their organization values them and what distinctive contribution they are expected to make in a world where AI handles the routine. Organizations that address this question honestly, by genuinely redesigning roles rather than simply adding AI tools to existing job descriptions, are far more likely to achieve the cultural alignment that makes transformation durable.

Where Catlin's Framework Invites Debate

The Rewired framework, compelling as it is, contains assumptions that deserve scrutiny.

The recommendation to focus on one to three domains is based on patterns observed in successful transformations. But selection bias is a genuine concern. The organizations that McKinsey works with on enterprise AI transformations are typically large, well-resourced companies with the luxury of making deliberate sequencing choices. For mid-market organizations facing competitive pressure from AI-native competitors who are deploying the technology broadly from inception, the luxury of a phased, focused approach may not be available. The competitive dynamics in some industries may punish deliberate focus just as surely as they punish unfocused sprawl.

Furthermore, the emphasis on process redesign before technology deployment, while conceptually sound, can become a justification for delay. Organizations that spend two years redesigning processes before implementing AI may find that the technology has advanced beyond the process architecture they designed for. There is a speed dimension to AI adoption that pure strategic discipline does not fully address.

These are not fatal objections to Catlin's framework. They are important refinements. The core insight—that focus, process-first thinking, and genuine organizational change are the determinants of AI transformation success—is supported by substantial evidence and practical experience. But leaders applying this framework should calibrate it to their specific competitive context rather than treating it as universally prescriptive.

What Business Leaders Should Do Now

For executives navigating AI transformation decisions, Catlin's framework suggests a practical set of priorities.

Catlin's central message is ultimately one about strategic maturity. The organizations that succeed with AI will not be those that treat it as a technology procurement exercise or a portfolio of innovation experiments. They will be those that treat it as what it genuinely is: a strategic transformation that requires the same discipline, clarity, and leadership commitment as any fundamental change in competitive positioning. The AI is not the hard part. The organization is.

References