Moving AI potential to performance

Almost overnight, AI has rewritten the rules of competitive advantage and created new opportunities to rethink how work drives performance. Now, leaders are under increasing pressure to deliver results.

Boards expect productivity gains. CEOs are focused on translating AI investment into measurable ROI. HR is being asked to transform potential into growth, often before there's full alignment on what should change.

The question is no longer simply how to adopt AI, but where to focus first to create meaningful, scalable value. Speed matters. Precision matters more.  

This starts with a deliberate approach to how work should change, where human judgment must remain and how AI can genuinely improve productivity, growth and competitive advantage.

The irony of AI transformation

The irony of AI transformation is this: The instinct to automate everything is exactly what gets in the way of creating value. But automation alone isn’t a strategy and rarely creates enterprise value. Value comes when organizations redesign work, reshape roles and build the skills and culture needed to sustain change.

AI transformation starts with the work itself. At the task level, leaders can see where AI can automate, where it can augment and where human judgment remains essential. That precision matters because work doesn't change evenly and, in some cases, AI creates entirely new tasks.  

Treating jobs as fixed units can lead to broad-brush workforce decisions. Instead of asking, “Which jobs should we eliminate?” organizations should ask, “Which parts of the work should evolve, and where will that create measurable value?” and “What specific tasks will be done by AI, a combination of AI and humans, and remain fully human-led, in order to focus job redesign?”

WTW’s analysis of more than 900 occupations in the Occupational Information Network reveals how unevenly distributed AI disruption is likely to be:

• Highly structured roles in operations, administration and clerical functions face 60% to 70% task automation potential
• Industrial and frontline roles could see up to 75% of repeatable tasks displaced
• Professional and judgment-intensive roles — where leaders often assume humans are less exposed — face 20% to 30% task reconfiguration

What we see consistently isn't uniform disruption or mass elimination but deliberate redesign of jobs at scale. Task-level precision moves AI activity to measurable value creation rather than simply layering technology onto unchanged work.

The work design challenge

Many organizations are learning the hard way: AI doesn’t just surface automation opportunities; it exposes where workflows can be improved and made more efficient. Unclear accountabilities, inconsistent role structures and a heavy dependence on tacit knowledge and human intervention are signs of fragmented workflows. AI won’t fix them. It scales them, reinforcing inconsistency and gaps, and adding complexity.

Let’s take this example: HR introduces AI to automate job architecture and skills mapping. Roles are generated quickly and skills mapped at scale, but underlying definitions are inconsistent, accountabilities overlap and critical knowledge sits with managers rather than in structured data. As outputs scale, problems compound. Roles are misclassified, and employees receive conflicting signals about expectations and career paths. The organization moves faster, but with less clarity. What looked like progress starts to erode alignment and accuracy.

Adoption is another factor organizations frequently underestimate. Even with a strong business case, change and value stall when leaders don’t communicate clear expectations for how work should change and how employees are expected to use AI.

Technology readiness isn’t the same as workforce readiness, and both need to be addressed. AI can accelerate execution, but it can’t create value when operating models are unclear or change capability is thin.

Getting this right means making decisions across a four-dimensional framework simultaneously:

• Define desired business outcomes  
• Identify work that will and will not change  
• Determine AI’s role across automation, augmentation or new value creation and where human judgment must remain and be built into workflows
• Establish a path for change acceleration to support AI transformation

Progress comes from sequencing work design deliberately, starting where value is high and friction is low. 

Where efficiency stops and value begins

If leaders focus only on short-term efficiency, they risk cutting capacity needed in other parts of the organization or to create long-term enterprise value.

The question isn’t, “Where can we cut?” but “Where should value shift?”

In one organization, administrative teams automated scheduling and reduced manual workload. But because there was no plan for redeployment, the capacity created simply sat idle. Efficiency creates capacity. What happens next determines whether that capacity becomes value. And in many organizations, that next step is still unclear.  

Value is only realized when leaders make disciplined decisions about where work moves, how roles are redesigned and which capabilities to build next. This includes redeployment, reskilling, targeted workforce investment and changes to the operating model. Without that follow-through, gains remain isolated, fail to scale and fall short of the durable competitive advantage that boards and investors expect.

Embedding AI the right way

Creating lasting impact through AI isn’t about slowing down; it’s about moving with greater precision. This starts with a fact-based view of how work actually gets done at the task level — where automation, augmentation and human judgment can be evaluated with clarity.  

This approach gives leaders a clearer view of how work needs to be redesigned and what to prioritize. It shifts the focus from broad enterprise-wide change to a targeted set of high-value moves, while balancing near-term efficiency with long-term workforce capability and change readiness.

When you align these elements and bring your people along with you, your organization is better positioned to scale value.  

Ultimately, competitive advantage in AI will come from decision precision, not automation volume. Work will change. But will it change by design or by default?