Individuals working with generative AI can now match the output quality of a two-person expert team. The more interesting question is why so many organizations still struggle to translate these individual gains into measurable business performance.

Technology Changes Capability. Systems Create Performance.

One of the assumptions I challenge most often when speaking with leadership teams is the idea that better AI tools automatically lead to better organizational performance.

They do not.

Generative AI is rapidly expanding what individuals can accomplish. Professionals can access knowledge outside their own area of expertise, solve more complex problems independently and complete demanding work in less time. Yet these individual gains do not automatically translate into better organizational outcomes.

The reason is straightforward. Organizations create value through systems, not individuals alone. When technology changes what people are capable of doing, roles, workflows and management practices need to evolve alongside it. Otherwise, much of the additional capability remains unused.

This idea sits at the heart of Human Performance Intelligence™. Performance emerges through the interaction between people, technology and the organizational environment. Changing one part of the system without adapting the others limits the value that organizations ultimately realize.

A recent Harvard Business School and NBER working paper offers compelling evidence for exactly this dynamic.

What Recent Research Shows

Researchers studied 776 professionals at Procter & Gamble as they worked on realistic business challenges under different conditions. Some worked individually, some collaborated in traditional cross-functional teams, and others completed the same tasks with access to GPT-4 or GPT-4o.

The findings were striking.

Individuals using AI produced work that matched, and in some cases slightly exceeded, the quality of work produced by traditional two-person expert teams. They also completed their work significantly faster. AI also reduced the divide between commercial and technical specialists by allowing individuals to access knowledge that would previously have required collaboration across functions.

The researchers observed another important effect. Participants reported more positive emotions and fewer negative emotions while working with AI, suggesting that the technology influenced not only performance but also the experience of solving complex problems.

Taken together, these findings suggest that generative AI does much more than automate routine work. It expands the cognitive resources available to an individual while making demanding knowledge work feel more manageable.

Looking Beyond the Results

The most interesting conclusion is not that AI allows individuals to perform at the level of traditional teams. The more significant insight is that AI fundamentally changes what one person can realistically contribute.

This distinction matters because increased capability and increased organizational performance are not the same thing.

A professional who can solve more complex problems independently still operates within an organizational system designed for yesterday’s assumptions. Reporting structures, approval processes, staffing models and performance expectations often remain unchanged, even though the nature of the work has shifted.

This helps explain why many organizations report enthusiastic AI adoption while struggling to generate meaningful business impact. Technology has advanced faster than the organizational systems designed to support it.

The Harvard findings reinforce an important proposition within Human Performance Intelligence™: technology creates new possibilities for human performance, but organizations capture those possibilities only when they redesign the environment in which people work.

The research also highlights something that deserves more attention. Participants did not simply produce better work. They experienced the work differently. Improvements in positive emotion suggest that better cognitive support can make demanding work feel less mentally taxing. As organizations increasingly depend on complex knowledge work, this relationship between cognitive performance and human experience is likely to become a growing source of competitive advantage.

What This Means for Leaders

For leadership teams, the implications extend well beyond selecting the right AI tools.

The first question is no longer whether AI improves individual performance. The evidence increasingly suggests that it does.

The more important questions are organizational.

Are roles still designed around assumptions that no longer hold?

Are managers redefining expectations as individual capability changes?

Are workflows allowing employees to make use of their expanded capacity?

Are performance measures capturing business outcomes rather than simply AI adoption?

Organizations that continue treating AI as a technology initiative are likely to realize incremental improvements. Organizations that redesign work around newly expanded human capability are far more likely to create lasting competitive advantage.

A Broader Lesson

One study rarely changes our understanding of work.

What makes this research valuable is that it aligns with a broader pattern emerging across many studies on AI and organizational performance. Individual capability is improving rapidly. Organizational adaptation is progressing much more slowly.

Human Performance Intelligence™ was developed to understand exactly this interaction. Rather than asking whether AI works, it asks a more useful question: under which human and organizational conditions does new capability become sustainable performance?

The answer is becoming increasingly clear.

Technology may expand human capability.

Organizations determine whether that capability becomes performance.