The New Productivity Divide

There has always been a divide between high-performing organisations and those that struggle to execute. That divide is typically explained through familiar lenses such as strategy, talent, leadership, or access to capital. Some organisations simply make better decisions. They move faster and operate with greater structural alignment than others. The differences are often apparent, and importantly, relatively stable. Improvement, while difficult, tends to follow recognisable paths.

What is now emerging is a different kind of divide — an AI productivity divide. This divide is beginning to reshape how work is performed and how value is created across organisations. It is less noticeable, less formally acknowledged, and significantly more volatile. It is not primarily driven by access to resources or even by differences in strategic intent. Instead, it is being shaped by how organisations, and more precisely the individuals within them, are engaging with artificial intelligence in their everyday work. This is closely tied to uneven AI adoption in the workplace. Capability is no longer evenly distributed, even within the same team.

At first glance, the concept of productivity may appear unchanged. Organisations continue to measure output, efficiency, and performance using familiar metrics. Leaders continue to speak in terms of optimisation, transformation programmes, and capability building. Yet beneath this surface, something fundamental is beginning to shift. This is productivity in the AI era. Productivity is no longer determined solely by individual effort, experience, or even organisational systems. It is increasingly influenced by an individual’s ability to extend their thinking, accelerate their execution, and amplify their output through AI-enabled tools.

This does not occur evenly across the organisation. AI does not wait for formal adoption programmes or enterprise-wide implementation strategies. It begins locally, often informally, and sometimes without explicit visibility. Individuals experiment. They integrate tools into their workflows. They find ways to reduce cognitive load, accelerate analysis, and produce work at a level of speed and quality that would previously have required significantly more time or support.

The result is not simply incremental improvement. It is divergence.
Within the same organisation, individuals performing similar roles can now operate at fundamentally different levels of productivity. One may complete a task in a fraction of the time, with greater depth of insight and clearer articulation, while another continues to rely on traditional methods. The difference is not always attributable to skill or experience. It is increasingly tied to how effectively each individual is leveraging AI as an extension of their capability.

This creates a form of digital productivity inequality that organisations are not yet structurally designed to recognise or manage.

Historically, productivity gaps could be addressed through training, process improvement, or system upgrades. The assumption was that capability could be standardised and scaled. In the current environment, that assumption breaks down. AI tools are widely accessible. The knowledge required to use them is unevenly distributed. More importantly, the willingness to integrate them into daily work varies significantly across individuals and teams.

The divide, therefore, is not only technical. It reflects a growing workforce capability divide that is behavioural and cognitive in nature.

Some individuals approach AI as a tool for augmentation. They actively explore its potential, refine how they use it, and embed it into their workflows in a way that compounds over time. Others remain cautious, uncertain, or disengaged. In some cases, this hesitation is rational, particularly in environments where governance, data security, and accountability are not clearly defined. In other cases, it reflects a deeper uncertainty about how work itself is changing.

From an organisational perspective, this creates a complex and often unobserved dynamic. Output begins to vary in ways that are difficult to explain. Certain teams move faster, produce higher-quality work, and appear more responsive. Others experience delays, bottlenecks, or increasing pressure. Leaders may attribute these differences to effort, capability, or management effectiveness, without fully recognising the underlying driver.

The reality is that productivity is becoming partially decoupled from traditional inputs.

On it’s own, effort is no longer a reliable predictor of output. Increasingly, decision-making efficiency through AI is becoming a defining factor in how quickly and effectively work moves from analysis to execution. Experience, while still valuable, does not guarantee efficiency. Even well-designed processes can be bypassed or accelerated by individuals who have found more effective ways of working. The organisation, as a system, begins to experience uneven acceleration.

This has implications not only for performance, but for how work is perceived and evaluated.

In environments where this divide is not acknowledged, a subtle but significant tension can emerge. High-output individuals may be seen as exceptional, without a clear understanding of how their performance is being achieved. Others may appear to underperform, despite operating within the same structural conditions. Over time, this can distort performance management, create misaligned expectations, and introduce questions around fairness and consistency.

More critically, it introduces risk.

If AI is influencing how analysis is conducted, how insights are generated, and how decisions are prepared, then it is also influencing the conditions under which decisions are made. Yet in many organisations, this influence remains largely invisible. There is limited clarity on where AI is being used, how it is shaping outputs, and what assumptions are embedded within those outputs.
This creates a disconnect between accountability and execution.

Leaders remain accountable for outcomes, but the processes through which those outcomes are being produced are becoming increasingly distributed and, in some cases, opaque. The productivity gains are real, but so is the potential for inconsistency, error, or misalignment.

At the same time, the opportunity is significant.

Organisations that recognise and engage with this shift have the potential to redefine what high performance looks like. The productivity divide is not inherently negative. It is a signal. It highlights where capability is being extended, where work is evolving, and where new forms of value are emerging.
The question is whether organisations choose to actively shape this divide, or allow it to develop unchecked.

Shaping it requires a different approach to productivity itself. Rather than viewing productivity as a function of output alone, it becomes necessary to consider how that output is produced. What tools are being used? How is thinking being augmented? Where is human judgement being applied, and where is it being deferred to systems?

This is not a purely technical exercise. It is a governance question.
If individuals are operating at different levels of augmented capability, then organisations need clarity on how that capability is integrated, validated, and aligned with decision-making processes. Without this, productivity gains may come at the expense of coherence.

There is also a cultural dimension.

The willingness to engage with AI is influenced by more than access or training. It is shaped by trust, incentives, and the signals that leadership sends out about what is acceptable and expected. In environments where experimentation is encouraged but boundaries are unclear, adoption may increase but consistency in how work and decisions are produced is no longer the same across individuals and teams. In environments where control is prioritised without enabling capability, adoption may stall altogether.
Navigating this requires balance.

Organisations need to create conditions where individuals can extend their productivity, while also maintaining clarity on accountability, quality, and alignment. This is not achieved through blanket policies or isolated training sessions. It requires an integrated view of how work is actually being performed.

It also requires visibility.

One of the defining characteristics of the current shift is that much of it happens below the surface. Individuals adopt tools independently. Workflows evolve without formal documentation. Capabilities develop unevenly across the organisation. Without mechanisms to surface and understand these changes, leadership is effectively operating with an incomplete view of how productivity is being generated.

This is where the concept of the productivity divide becomes useful.
It provides a lens through which to observe what is already happening. It shifts the conversation from whether AI will impact productivity to how that impact is currently being distributed. It also creates a basis for more targeted intervention.

Rather than attempting to drive uniform adoption, organisations can focus on identifying where productivity is being significantly enhanced, understanding the mechanisms behind it, and determining how those mechanisms can be responsibly extended.

This approach recognises that the future of work is unlikely to be uniform.
Variation will exist. Individuals will operate at different levels of augmented capability. The role of the organisation is not to eliminate that variation entirely, but to ensure that it does not undermine coherence, accountability, or strategic direction.

Ultimately, the new productivity divide is not about technology alone. It is about how technology interacts with human capability, organisational structure, and decision-making processes.

It challenges long-standing assumptions about what drives performance.
It raises questions about how productivity should be measured, how capability should be developed, and how organisations maintain alignment in an environment where work is increasingly fluid and distributed.

Most importantly, it introduces a degree of urgency.

This is not a future scenario. The AI productivity divide is already forming, shaping the reality of future of work productivity in ways many organisations have yet to fully see. In many organisations, it is already present. The risk is not that it will emerge, but that it will go unrecognised until its effects become difficult to manage.

For senior leaders, the implication is clear.

The question is no longer whether AI will influence productivity. It is whether you have visibility over how that influence is already reshaping your organisation, and whether you are positioned to respond.

Because in this environment, productivity is no longer just a measure of output.

It is a measure of how effectively capability is being extended, integrated, and governed.

And that is where the real divide now operates.