What if the biggest competitive advantage in the age of AI isn’t AI?

Organizations are investing billions in increasingly sophisticated technology. They are building AI strategies, experimenting with agents, introducing copilots, training employees to prompt, and racing to embed AI into workflows. All of this makes sense. But there is a question I think we should be asking more often:

What happens when everyone has access to increasingly similar intelligence?

When your competitors can access the same foundation models, generate similar analyses, summarize the same market information, and automate many of the same processes, where does differentiation come from? Perhaps the organizations that benefit most from AI won’t necessarily be those with the best AI. Perhaps they will be the ones with the people who know what to do with it.

When intelligence becomes abundant, something else becomes scarce

For most of business history, access to expertise was a competitive advantage. Organizations hired people because they knew things others didn’t. Experience accumulated over decades mattered because knowledge was difficult to acquire, interpret and transfer.

AI is beginning to change that equation.

Information that once took hours or days to find can now be synthesized in seconds. AI can generate strategic alternatives, analyze large datasets, draft communications, challenge assumptions, identify patterns and increasingly support complex decisions. The implications are enormous. But they also create a paradox.

As machine intelligence becomes more powerful and more widely available, access to intelligence itself may become less differentiating. The differentiator moves elsewhere.

• It moves toward the ability to ask the right question.
• To recognize when an answer doesn’t make sense.
• To understand context the machine cannot see.
• To challenge assumptions.
• To make trade-offs when there is no objectively correct answer.
• And ultimately, to decide the best course of action that will create a competitive advantage

In other words, as artificial intelligence becomes more abundant, human judgment becomes more valuable.

AI can improve performance. It can also reduce it.

One of the most interesting studies on this question comes from researchers working with Boston Consulting Group. In a large field experiment involving 758 consultants, researchers examined what happened when professionals used GPT-4 to complete realistic consulting tasks. The results were impressive. For tasks that fell within what the researchers called AI’s “technological frontier,” consultants using AI completed 12.2% more tasks and worked 25.1% faster. Their work quality also improved significantly.

But then something interesting happened. When consultants were given a complex task deliberately designed to sit outside AI’s capabilities, those using AI were 19 percentage points less likely to arrive at the correct solution than those working without it. The researchers described AI’s capabilities as a “jagged technological frontier.” AI can be extraordinarily capable in one area and surprisingly unreliable in another, even when the tasks appear similar.

That raises an important leadership question. If AI itself cannot reliably tell us where its frontier lies, who decides when to trust it? Navigating the frontier becomes a leadership capability. The challenge is that AI doesn’t arrive with a warning light telling us when we have crossed its frontier. A confident answer inside the frontier can look remarkably like a confident answer outside it. So how will leaders know? Perhaps they won’t always know. Instead, they will need to develop the judgment to recognize when greater human scrutiny is required.

The question changes from “Can AI do this task?” to a more sophisticated set of questions:

• Is the answer verifiable?
• What context might AI be missing?
• What, if any, are the conflicting signals?
• How consequential is the decision?
• And do we have enough human expertise in the room to recognize when the answer is wrong?
• This suggests an emerging leadership capability that goes beyond AI literacy:

AI discernment, the ability to decide when to delegate to AI, when to collaborate with it, when to challenge it, and when to override it.

And because the technological frontier will keep moving, this cannot become a static set of rules. Leaders and organizations will need to keep experimenting, learning, and redrawing the boundary. That may ultimately be where competitive advantage lies, not simply in having more powerful AI, but in knowing where AI ends, and human judgment needs to begin.

Human + AI does not automatically equal better

Another assumption is worth challenging: that combining human intelligence with artificial intelligence will naturally produce superior results. The research suggests otherwise. A 2024 meta-analysis published in Nature Human Behavior examined 106 experiments and 370 effect sizes comparing humans working alone, AI working alone, and humans working with AI. Human-AI combinations generally performed better than humans alone. Human-AI combinations generally outperformed humans alone. But surprisingly, when either the human or the AI was already better at a particular task, combining them often reduced the stronger one’s performance. The effect was particularly striking in decision-making tasks, where combining humans and AI produced significant performance losses.

The researchers found more promising results for creative tasks, but their broader conclusion is important: simply putting a human and an AI together does not guarantee synergy. Think about what that means for organizations. Buying powerful AI is not enough. Giving everyone access to it is not enough. Training people to prompt it is not enough.

The real organizational capability is knowing how human and artificial intelligence should work together.

• When should AI lead?
• When should the human lead?
• When should AI generate possibilities while humans evaluate them?
• When should people challenge the machine?
• And when should they ignore it completely?

Those are not primarily technology questions. They are questions of judgment and work design. The competitive advantage may lie in the decisions after the answer. Imagine two competing organizations. Both have access to highly capable AI. Both ask AI to analyze customer trends, competitor activity, and market data. Both receive broadly similar insights. One organization takes the recommendations largely at face value and moves quickly. The other asks different questions.

• What assumptions sit underneath this analysis?
• What information might be missing?
• What does the AI know about our customers, and what doesn’t it know?
• How does this recommendation fit our strategy?
• What would have to be true for this conclusion to be wrong?
• What second-order consequences might we be overlooking?
• What are we seeing in the market that isn’t yet visible in the historical data?

The competitive advantage isn’t necessarily in the quality of the first answer. It is in the quality of the thinking that follows it. This may become increasingly important as AI outputs become more polished and persuasive. A poorly written recommendation invites scrutiny. A beautifully structured, confident and data-rich recommendation can create the opposite reaction. It can make us less likely to question it.

That makes critical thinking and judgment not less important in an AI-enabled organization, but more important. To stay competitive, adaptability and going a layer deeper may become equally important. The World Economic Forum’s Future of Jobs Report 2025 found that analytical thinking remains employers’ most important core skill, essential to seven out of ten companies. Immediately behind it are resilience, flexibility, and agility, followed by leadership and social influence. When the World Economic Forum looked at the skills distinguishing growing roles from declining ones, resilience, flexibility, and agility emerged as the strongest differentiator.

That matters because AI is not simply changing individual tasks. It is changing work itself. Roles are evolving. Decision rights will move. Expertise will be redefined. Some activities will disappear while entirely new ones emerge. The people who thrive in this environment will not simply be those who master today’s AI tools. Today’s tools will inevitably change.

They will be people who can continuously rethink how they create value.

That requires curiosity, learning agility, and the willingness to experiment. And perhaps most difficult of all, the willingness to let go of expertise that once defined our professional identity.

The technology may be easier to copy than the organization

Evidence already shows that simply adopting AI does not automatically translate into enterprise value. McKinsey’s 2025 global AI research found that more than three-quarters of respondents said their organizations were using AI in at least one business function. Yet more than 80% said they were not seeing tangible enterprise-level EBIT impact from generative AI. One organizational factor most strongly associated with financial impact was not having a particular model. It was redesigning workflows.

That distinction matters because technology can be purchased, models can be accessed, tools can be copied. But an organization’s collective ability to rethink work, make good decisions, challenge assumptions, learn quickly, and adapt its operating model is much harder to replicate. AI lets us do things faster, but we need to focus on doing the right things. This is where leadership becomes critical.

The leadership challenge is no longer simply: How do we get people to use AI?

It is: How do we create an organization that thinks better because it has AI? Those are very different ambitions.

From AI adoption to Human + AI capability

Perhaps we are measuring the wrong things. Many organizations are understandably tracking AI adoption:

• How many employees are using the tools?
• How many prompts are being submitted?
• How much time is being saved?
• How many processes have been automated?

These measures tell us whether AI is being used. They don’t necessarily tell us whether the organization is becoming better. We may need another set of questions.

• How are people making better decisions?
• How are they asking better questions?
• How quickly are they identifying assumptions?
• How are they exploring more alternatives before committing?
• How are teams learning faster?
• How are leaders using the capacity AI creates for higher-value thinking?

And perhaps most importantly:

That last question may prove particularly important. Because if AI consistently does the first draft, conducts the analysis, structures the argument and proposes the recommendation, we need to consider what happens to the human capabilities that previously developed through doing that work. Efficiency matters but capability matters too.

The next competitive advantage

We are entering a period in which extraordinarily powerful intelligence will become embedded in almost every organization. Of course, differences will remain in data, infrastructure, models, and implementation. But the technology itself is unlikely to be the whole story. The deeper differentiator may be the human system surrounding it. The organizations that outperform may be those whose people know when to trust AI and when to question it.

• Who can move quickly without surrendering judgment.
• Who can use AI to expand their thinking rather than replace it.
• Who can adapt as the technology changes.

And who understand that the purpose of AI is not simply to produce more answers. It helps us make better choices, uncover customer needs, and solve today’s pressing problems. So perhaps the question leaders should be asking isn’t: How intelligent is our AI?

Perhaps it is: How intelligent is our organization becoming because of it?

And if every competitor eventually has access to powerful AI, what will our people be able to do with it that others cannot?

Building Leadership Capabilities

The implication for organizations is significant. If human-AI collaboration does not automatically improve performance, simply giving leaders access to AI and teaching them how to use the tools is not enough. We need to develop leaders who understand where AI creates value, where human judgment must remain central, and how to design the interaction between the two. That means knowing when to delegate to AI, when to challenge its conclusions, when to bring in human expertise, and when not to use AI at all. The next phase of AI capability building therefore needs to move beyond adoption and prompting toward something much more important: developing leaders who can make intelligent choices about how AI is used to create value across the organization.

But developing this capability will require more than another skills training program. Organizations need to create regular dialogue among leaders to share experiments, failures, and lessons learned, using their collective wisdom to continually redefine where AI should and should not be used. Effective AI leadership should become part of performance expectations and rewards, recognizing leaders not simply for adoption, but for measurable value created and the quality of decisions made. Leaders also need opportunities to experiment with AI on real business challenges, with permission to test, learn, and challenge the technology. And as we invest in AI capability, we need to invest just as deliberately in the critical thinking, adaptability, and judgment required to use it well.

Perhaps, then, the goal should not be to create organizations where leaders use AI the most. It should be to create organizations where leaders know how to use AI wisely, combining the best of artificial intelligence with the best of human intelligence to create value that neither could achieve alone.