Jesse Jacoby
A few weeks ago I sat in on a strategy session where a senior leader pulled up a polished AI-generated analysis of his competitive landscape. It was thorough. The frameworks were sound. The data points were current. Around the table, heads nodded. The conversation moved quickly toward implementation.
Nobody asked whether the analysis had framed the right problem.
That moment has stayed with me, because it captures something I have been watching unfold across boardrooms, executive teams, and C-suites for the past two years. The bottleneck in business decision-making is no longer access to information or speed of analysis. Both are essentially free now. What has gotten rarer, and therefore more valuable, is the capacity to look at a confident, well-formatted answer and ask whether it actually matters.
Call it discernment. Call it judgment. Call it the willingness to stay with a question instead of grabbing the first plausible response. Whatever you call it, it is becoming the scarcest resource in modern leadership, and most organizations are unprepared for what that means.
The Productivity Trap
For thirty years, executive performance has been measured roughly by throughput. How many decisions made, meetings held, emails answered, initiatives launched, problems closed. AI has poured rocket fuel on that metric. A leader can now produce in an afternoon what used to take a team a week.
The trouble is that volume of output and quality of judgment have always been loosely coupled, and AI is loosening the coupling further. The leaders I see thriving with these tools are not the ones who use them to move faster. They are the ones who use the time AI gives them to think harder. They take the analyst-quality memo their tool produces and they slow down. They ask what is missing. They ask who would disagree, and why. They notice when an answer feels too clean.
The leaders I see struggling, often without realizing it, are using AI to skip the discomfort of not knowing. They mistake the speed of arrival at an answer for confidence in the answer. They confuse a well-written summary for a well-considered position.
This is not a tooling problem. It is an interior one.
Pattern Recognition Is Not Wisdom
Here is something the consulting industry does not say out loud: most of what gets called strategic thinking is pattern recognition. We have seen this before. Here is what worked. Here is the framework that fits.
AI is preposterously good at pattern recognition. It will, with breathtaking speed, find every analogous situation in the corpus of human business writing and surface the pattern that fits your problem. This is genuinely useful. It is also seductive. Because the pattern that fits is not always the pattern that is true.
Discernment is the ability to notice when the situation in front of you is not the situation the pattern describes. It is the small voice that says the data is right but the conclusion is wrong. It is the executive who sits with a unanimous recommendation and says, “I hear you. I am not sure yet.” It is a capability that does not improve with more information. Past a certain point, more information makes it worse.
The leaders who develop this capability tend to share a few traits that do not show up on competency models. They are comfortable being the slowest person in the room. They do not need to be the smartest. They have an unusual relationship with their own confidence, in that they distrust it slightly. And they ask questions that other people find inefficient.
The Approval Economy
A subtle dynamic is at work in many high-performing organizations. The very best people, the ones we promote and reward, are often the most approval-oriented. They got where they are by giving good answers, fast. They are wired to deliver. AI rewards this wiring. Ask a question, get an answer, ship it.
The problem is that approval-seeking and discernment are nearly opposite competencies. To exercise judgment, a leader has to be willing to be wrong publicly, to say “I do not know” in front of a board, to push back on a polished deliverable that has obvious flaws no one wants to name. The instinct that built their career, the instinct to deliver what is expected, frictionlessly, is the same instinct that prevents them from doing the harder work AI now makes possible.
I have started watching for this in coaching conversations. The tell is usually how a leader responds to ambiguity. Some leaders, when faced with a genuinely hard question, get visibly uncomfortable and pivot quickly to action. Others sit with it. The first group is impressive at meetings. The second group makes the calls that age well.
There is a quiet pathology here that we should be willing to name. A great deal of what gets celebrated as decisive leadership is actually anxious leadership. The pace was always a tell. Now AI lets the anxious leader move even faster, with even more polish, and the underlying lack of judgment is masked even more thoroughly. Until it is not.
What an Organization Built for Judgment Actually Looks Like
It is tempting, having identified discernment as a capability, to try to install it the way we install other capabilities. Training programs. Frameworks. Maturity models. None of these will work, and most of them will make things worse, because discernment is not a skill that responds to instruction. It is a practice that responds to conditions.
Some of the conditions I have seen actually work:
Slowness as a strategic choice. Some teams I work with have started building deliberate friction into their decision processes for high-stakes calls. Not bureaucratic slowness, but reflective slowness. A 24-hour rule before committing. A required dissenter. A scheduled pause between recommendation and decision. These feel inefficient because they are. That is the point.
Asymmetric reward for good questions. Most performance systems reward people for having answers. A few sophisticated organizations have started rewarding people for having questions that change the trajectory of a conversation. This is hard to measure, which is why almost no one does it, and which is why those who do gain a quiet advantage.
Senior leaders modeling uncertainty in public. The single most powerful thing an executive can do to develop organizational judgment is to say, in a meeting where doing so feels risky, “I do not know what we should do here. Let me think about it.” Most do not. The ones who do create permission for the rest of the organization to think rather than perform.
Protecting time that produces no output. Reflection, walking, journaling, a long conversation with a thoughtful peer, reading something irrelevant to the immediate problem. The activities that produce judgment look indistinguishable from leisure on a calendar, which is why they get cut first when the calendar gets full. AI is going to make calendars fuller. Leaders who do not actively defend unproductive time are about to find themselves with sharper tools and duller minds.
The Provocation
Here is what I would offer for any leader reading this who suspects they might have a discernment problem.
The next time you receive an analysis, a recommendation, a polished AI-generated brief, sit with it for an extra five minutes before you react. Do not edit it. Do not approve it. Do not push back. Just notice what your mind wants to do with it. If your first impulse is to act, that is information. If your first impulse is to validate, that is information. If your first impulse is to question what you actually think about the underlying problem, congratulations. You may already be in the small group whose value will compound over the next decade.
The leaders who thrive in an AI-saturated world will not be the ones who process more, faster. The cost of doing that has gone to zero. They will be the ones who can hold an answer at arm’s length long enough to ask whether it is the right one.
That capability is not new. But for the first time in a long time, it is rare. And rare, in markets, is where the value lives.
