McKinsey’s 2025 State of AI research found that 88% of organisations are now using intelligent systems in at least one function. BCG’s global survey of 1,250 senior executives, conducted the same year, found that only 5% qualify as high performers; that is, organisations generating material, measurable returns from their investment. The technology available to the 5% is, broadly speaking, the same as the technology available to the other 95%. What is different is not what they have access to. It is the quality of the leadership decisions that determine where the capability is applied, how seriously it is pursued, and whether the organisation is actually structured to capture the value it creates.
BCG’s research found that C-level executives who are deeply engaged with their strategy are 12 times more likely to be among the top 5% of value-creating organisations. Twelve times. The differentiator is not the tool. It is the judgement of the persons holding it.
What The Problem Actually Is
Most organisations are not failing because they lack access or awareness. They have both. They are failing because the quality of thinking applied to their deployment decisions has not kept pace with the ambition of what they are attempting. PwC’s 2026 analysis of the organisations struggling to convert investment into outcomes identifies a consistent pattern. Many take a ground-up approach by crowdsourcing initiatives, encouraging experimentation across functions, then attempting to shape the resulting activity into something that resembles a strategy. The outcome is projects that rarely match enterprise priorities, that are rarely executed with precision, and almost never produce the kind of transformation that justifies the investment. Impressive adoption numbers with limited outcomes.
McKinsey’s research supports this. Only 39% of organisations demonstrate any measurable financial impact from their investments, and for most of those, the impact represents less than 5% of EBIT. The gap is not a technology gap. It is the absence of leadership judgement that determines which problems are worth solving, which initiatives deserve concentrated investment, and which should be declined regardless of how technically capable the underlying capability appears.
What the organisations pulling ahead have understood is that the leaders winning are those who have stopped treating this as a technology race and started treating it as a management problem. They are rebuilding how decisions get made, how accountability is structured, and how the organisation determines what is worth pursuing.
What Judgement Actually Means in Practice
Judgement in leadership is a practice, and it operates through three specific capabilities.
The first is the ability to identify which problems are worth solving within a specific organisation’s context. The question every leadership team should be answering is not “where can we use this?” It is “where, in the specific way our organisation creates value, does this create an irreplaceable advantage?” That question requires deep operational knowledge of how the business actually works.
The second is the ability to connect capability to strategy, and to sustain that connection under pressure. PwC’s analysis of the organisations generating real returns identifies concentration and discipline as the defining characteristics: precision in selecting a small number of areas where the investment can deliver genuine transformation, then executing with sustained focus through the periods when results are not yet visible. Not breadth. Not the volume of active pilots. Concentrated strategic intent, held when the temptation to scatter resources across more initiatives is at its highest.
The third is the ability to ask the right question first. Most organisations start in the wrong place by asking what the technology can do rather than what the organisation needs to do differently. Starting with the technology produces a solution looking for a problem. Starting with the organisation’s most consequential unsolved problems produces a clear brief against which any capability can be honestly evaluated.
The Hardest Judgement Call
Of all the leadership decisions AI leadership requires, the hardest is the deliberate ‘no’. Not the no to an obviously irrelevant initiative. The hard no is the one said to an initiative that is technically capable, enthusiastically supported within the organisation, and genuinely interesting, but not aligned to the strategic priorities the leadership team has committed to. It is the no that comes after someone has invested considerable energy in building a case. The no that will be received as overcaution rather than clarity. This is where most organisations fail. Not from lack of aspiration, but from the inability to protect strategic concentration against the constant pressure to add more initiatives, respond to more trends, and appear to be keeping pace with every development in the external environment.
BCG’s data on the organisations generating material returns is unambiguous on this point. High performers pursue roughly half as many initiatives as the organisations still circling in pilot mode, and scale significantly more of them. The concentration is not caution. It is the direct product of leadership judgement applied to the question of what the organisation can actually absorb, execute with sustained discipline, and derive lasting value from. That is harder to exercise than the yes that keeps everyone in the room comfortable. It has to be made against the current. And it has to be made clearly, with a shared understanding of why.
The question that will determine whether an organisation captures genuine value from the AI shift is not how much capability it has access to. The question is how clearly the leaders of that organisation are thinking about where the capability creates authentic strategic value, and how disciplined they are about pursuing only that, consistently and over time. That clarity is not produced by expanding the portfolio of experiments. It is produced by the willingness to ask hard questions about what the organisation actually needs, about where it genuinely creates value, about what it can realistically absorb, and to hold the answers under real scrutiny rather than validating what is already under way.
It requires the same qualities that have always separated effective leadership from reactive management. The capacity to think clearly under pressure. The discipline to act on what that thinking produces, even when the action is to stop rather than start. And the confidence to hold a small number of strategic positions with conviction while an environment designed to undermine that conviction continues to do its work.
