AI has brought about an organizational reckoning. Change is inevitable, but digital change is unrelenting. AI rapidly accelerates the speed at which companies can and must operate. Many of the organizations that once led disruption are now being disrupted themselves. Change is no longer a temporary phase to manage. It is the constant operating condition in which modern organizations must function. 

Market signals now move so quickly that traditional organizational structures simply can’t keep up. Customer expectations are evolving at a phenomenal pace, influenced less by direct competitors and more by standout experiences encountered elsewhere. New rivals appear suddenly, often from unexpected sectors, unencumbered by outdated systems or legacy thinking. 

In today’s business world, strategy ages rapidly. Plans don’t fail because they were poorly designed; they fail because the landscape they were built for no longer exists. 

Leading strategy experts increasingly argue that in today’s dynamic business environment, competitive advantage is often temporary, requiring leaders to rethink the assumptions behind traditional long-term planning. 

Research from Boston Consulting Group on adaptive strategy suggests that in unpredictable and rapidly changing markets, advantage tends to be serial rather than sustained, i.e., businesses capture value for a period of time before shifts in technology, customer behaviour, or competitive structure force repositioning. 

While this pattern has been emerging for years, it has become more pronounced following the rise of agentic AI, which accelerates experimentation, compresses response times, and lowers the cost of strategic iteration across many industries. 

The companies that will endure over the next decade will not be defined by superior algorithms or larger datasets, but by the speed at which companies identify signals, make decisions, and execute3, by their willingness to fundamentally rethink how decisions are made, how leadership roles evolve, and how the business itself operates using human and digital labour.

Under constant pressure, most organizations respond by adding. More processes. More reporting. More tools. More layers of oversight. Each addition makes sense on its own, but collectively they create congestion. Decision cycles slow down. Accountability becomes blurred. Leaders spend more time navigating internal complexity than reading what is happening outside the organization. 

One place this accumulation shows up quickly is in calendars. As uncertainty increases, organizations add meetings to manage risk, align stakeholders, or show oversight. Over time, decision time shrinks while coordination expands. In 2023, Shopify addressed this by cancelling most recurring meetings across the company and setting limits on when new meetings could be added. The goal was to give people uninterrupted time to think, decide, and execute. Shopify acknowledged that meetings themselves had become a source of friction and that, without deliberate removal, attention would continue to fragment. 

As complexity grows, the instinct is to control it with more structure. Yet beyond a certain point, that structure begins to consume the very capacity it was meant to protect. It also starts to distort how reality is seen inside the organization. 

As information moves upward through layers of management, it is filtered, interpreted, softened, and repackaged until the signals that matter most are often the ones least likely to survive. By the time decisions reach senior leadership, what remains can look reassuringly polished but dangerously incomplete, creating a false sense of control built more on performance management than on truth5. 

To address this, organizations need to take things away rather than continue adding more. Strategic subtraction is the deliberate practice of removing layers, demands, and decision burdens that diminish clarity, judgment, and execution. It is not minimalism for its own sake. It is an approach grounded in a simple reality many leadership teams prefer not to confront: cognitive capacity and attention are finite, and decision quality does not improve simply because more information is circulated. 

When those limits are exceeded, performance rarely breaks down dramatically. It erodes quietly. Meetings multiply, reactive decision-making becomes more common, and metrics expand in ways that create activity without producing much insight. The organization does not suddenly collapse. It slowly drifts as accumulation places a growing strain on the system, making effective execution harder than it should be. That is what makes accumulation so dangerous. It rarely presents itself as a crisis. It simply taxes the organization over time until everything feels heavier, slower, and less clear. 

Subtractive leadership starts from a different question. Not what else do we need, but what no longer earns its place. Which processes remain only because removing them would be politically uncomfortable? Which reports are produced out of habit rather than necessity? 

Which tools address problems that are no longer strategic? These questions are difficult because subtraction creates visible consequences. Adding rarely does. 

This is also why subtraction is resisted. Legacy processes almost always have an owner. Past decisions develop defenders. Risk functions are rewarded for preventing removal, not enabling it. Over time, organizations accumulate protective layers that once served a purpose but now restrict movement. The friction is internal, not technical. 

Amazon addressed this problem by deliberately constraining team size through its Two-Pizza Teams model. Teams were designed to be small enough to be fed by two pizzas, typically five to eight people, with clear ownership and authority to make decisions independently. The objective was not to move faster at any cost, but to reduce coordination overhead, limit approval chains, and prevent internal complexity from crowding out judgment. By shrinking the unit of decision-making, Amazon reduced the need for status reporting and cross-team dependency, preserving clarity as the organization scaled. 

Strategic subtraction forces leadership teams to confront which activities, roles, or decisions matter less than they once did. That choice is rarely stated outright, but it is felt immediately. Faced with that tension, many organizations choose manageable inefficiency over visible conflict. And so, the accumulation continues. 

Automation is often presented as a means to simplify organizational complexity. Yet, in reality, it doesn’t simplify – it exposes. When processes become automated, previously hidden issues become glaringly obvious. Informal workarounds and implicit understandings must now be explicitly defined. Any gaps in ownership or data quality suddenly surface. Rather than streamlining operations immediately, automation reveals long-masked inefficiencies, inconsistencies, and ambiguity. 

This exposure often stalls automation initiatives. It’s not because the technology fails but because the organization is unprepared for the transparency automation demands. Procurement processes slow down as requirements shift. Compliance 

teams raise last-minute concerns. Business units defend their unique ways of working as strategically essential. The promised efficiencies don’t materialize because the organization wasn’t ready to face its internal complexity with an open mind. 

When implemented effectively, automation’s real benefit isn’t just reducing costs. It’s about consistency. Tasks become standardized, reducing errors and variability. This consistency forms a reliable operational baseline, enabling organizations to generate meaningful insights. Without consistent processes, analytics, and strategic decisions, the organization remains unstable and unreliable. 

McKinsey’s Global Fashion Index tracks performance across 350+ publicly traded fashion companies, and the latest iteration makes a blunt point about what separates leaders from the pack. The leaderboard is dominated by retailers that moved early on store-level inventory accuracy and adopted the enabling technology, RFID, needed to get there. In some sectors, the first real advantage comes from doing the unglamorous basics well: making inventory data accurate enough that forecasting, replenishment, and fulfillment aren’t built on assumptions. 

Manufacturing offers a long-running illustration of the same principle. Toyota’s production system emphasizes continuous improvement, waste reduction, and decision-making at the lowest responsible level. Rather than relying on centralized oversight, teams are expected to surface problems early, correct them locally, and refine processes incrementally. This design has allowed Toyota to respond more effectively to supply chain disruptions and market shifts than organizations dependent on rigid planning cycles or centralized control. The system works because internal friction is deliberately reduced, preserving the organization’s ability to act clearly as conditions change. 

With consistent automation, leaders begin to see deviations as meaningful signals rather than routine noise. Attention shifts to important, value-added activities rather than constantly managing and policing repetitive tasks. 

Artificial intelligence amplifies the benefits of automation but introduces new challenges. AI models excel at identifying subtle patterns within vast data sets, surfacing insights that human analysis typically overlooks. However, rather than simplifying decision-making, this flood of new insights can overwhelm leaders, causing interpretive overload. 

Leaders find themselves confronted with more information than they can comfortably handle. AI-generated recommendations can seem technically sound yet practically ambiguous, clashing with leaders’ intuition shaped over years of experience. Trust in AI becomes an issue, not because leaders distrust the technology itself, but because they worry about the consequences of following its recommendations within their organizations. 

To overcome this gap, organizations need to connect analysis directly to decision-making through Decision Intelligence. Decision Intelligence moves beyond producing insights; it focuses on clearly and explicitly understanding the consequences of choices. Questions shift from “What will happen?” to “What decisions must we make if this happens?” Leaders start discussing explicit assumptions, trade-offs, and tolerances rather than abstract predictions. 

When Decision Intelligence is implemented successfully, organizational debates become clearer. Disagreements don’t vanish but become more productive and precise. Leaders focus discussions on clearly defined inputs and assumptions rather than vague, subjective outcomes, fostering a stronger culture of organizational learning. 

A critical barrier to effective AI adoption is organizational alignment. Decision Intelligence challenges traditional power structures. Leaders accustomed to controlling decisions through exclusive access to information must now justify their choices openly. Employees who previously had to escalate decisions upward must now manage trade-offs themselves. Redistributing cognitive effort across the organization can feel threatening, especially where authority traditionally comes from withholding information. 

For example, logistics companies have struggled to implement advanced optimization systems successfully. Regional leaders often override AI recommendations when they conflict with local incentives. Only after clearly defining incentives and aligning decision-making criteria across the organization does AI begin to deliver its promised value. 

Furthermore, most organizations are not limited by a lack of intelligence. They struggle with ambiguity. Broadly stated strategies maintain internal consensus, overlapping roles ensure flexibility, and unclear decision rights avoid internal conflict. Yet these ambiguities slow down decision-making and execution. 

Strategic subtraction tackles this by forcing clarity. It demands specificity and reduces interpretive confusion. When employees know exactly what’s expected, they act confidently and rapidly. When roles and decisions are ambiguous, caution and hesitation dominate.