Early in my consulting career, I built a business case for an eProcurement system. No AI, just me, a spreadsheet and more late nights than I would like to admit. I did the research, built the model line by line and wrote down every assumption. Before it went to the client, my senior manager reviewed my work and asked question after question.
Then came the presentation to the CFO. She asked one question: “How will the ROI be realized?” She wasn’t asking about the spreadsheet. She was testing whether the plan was realistic. I walked her through the thinking behind the numbers, well beyond what was on the page. I could do that because I had done the work and because my manager’s questions had prepared me for hers.
Fast-forward thirty years. Whether I am mentoring young professionals or catching up with colleagues from my consulting days, we still tell those stories: the all-nighters to prepare for a facilitated session, writing and rewriting executive communications until the tone and positioning were right, the deck rebuilt at 2 a.m. We laugh about them, but the lessons still ring true. That was where we learned how to work, and how to think critically. Doing the work, and having a manager review it, were the bottom two rungs of the career ladder.
Now picture this. A young analyst asks AI to build the same business case. In an hour, she has clean slides, a tidy model and a confident ROI. No one reviews the model or the assumptions. She presents it, and the CFO asks the same question: “How will the ROI be realized?”
Silence. The AI assumed it. She never checked, and if she’s honest, she wouldn’t have known how.
The problem isn’t that she used AI. Every professional, at every level should. The problem is that she trusted it and no one reviewed her work.
The work that taught us how to work is disappearing. McKinsey lists the tasks AI is absorbing: research, documentation, data cleanup, basic coding, preliminary analysis. The list goes on. That was the training ground. Slogging through those assignments is how we learned what good looks like, and it’s what lets us challenge AI today. If you have never done the work, how do you know when AI is wrong? How do you know if you wrote a “good” prompt?
AI has raised the bar: young professionals are expected to deliver more, sooner, with less of the on-the-job learning that builds business chops. As Fast Company put it, AI can speed up early-career performance, “but it cannot accelerate the accumulation of judgment.”
The analyst wasn’t the only one who skipped a step. Her manager did too. My manager’s questions were half my education, and they are what prepared me for the CFO’s.
Today, young professionals tell me their managers rarely make time to review work. Teams are leaner, calendars are full, and remote work has ended the side-by-side review. And to boot, AI-generated work looks finished, even when the thinking isn’t. In a Harvard and BCG study cited by Fast Company, consultants using AI produced more polished work, yet junior consultants struggled to explain where their recommendations might break down. Polish is not proof of rigor.
That changes part of the manager’s job. Repetition used to build judgment. Now AI absorbs the repetition, so judgment has to be built deliberately, and the review is the main place left to build it. AI can’t do that review for you. It can check the formulas, but it can’t explain the business context behind the answer or connect the dots that make an analyst better. Skip the review, and we remove the last rung. The cost won’t show up in the next deck. It will show up years later, when we promote people into roles that demand instincts no one helped them form. If we don’t build those instincts now, when will we?
If you are early in your career, learn to do the work, use AI to be efficient, and consider the following questions before submitting your work.
Remember, AI doesn’t replace experience, human judgment and context.
If you manage young professionals, you are the answer to question five above. As managers, we play an integral role in shaping future leaders.
To this day, I remember the tough questions asked and the learning moments. When the next executive asks a question, the answer shouldn’t depend on who wrote the prompt. It should depend on who did the work, and who took the time to review it.
Stay curious. Learn to do the work. After all, the company hired you, not the machine.
Every organization runs on two tracks: the one on paper, and the one in the room.
The second track has no agenda. It runs on history that adds weight to every conversation, politics that decide whether a good idea gets oxygen or quietly disappears, and power dynamics that move faster than any org chart.
AI has made the first track easier than ever to analyze. The second still requires human judgment: deciding what matters, how to move, and what trade-offs are worth making. Professionals increasingly turn to AI to work through high-stakes situations, but there are signals AI cannot read, rooms it has never been in, and consequences it cannot weigh.
That space is the judgment gap: the widening space between intelligent information and the contextual wisdom required to act. It shows up in every room where decisions land.
Deloitte’s State of AI in the Enterprise 2026 puts it painly: advanced organizations are redesigning work so AI executes routine work efforts, while humans focus on judgment, exception handling, and strategic oversight. Overtime, we should expect to see roles redesigned with a complimentary AI workforce and playing to both the human and AI strengths.
Three Places the Gap Shows Up
Inside organizations. Teams are often blind to the drag in their own system. Leaders cannot fix what they cannot see, and those growing into their authority lack the pattern recognition to challenge what AI tells them. Distance reveals what proximity hides.
Inside careers. A recent Harris Poll found two-thirds of adults aged 18–25 feel they are missing opportunities because they lack mentors to help them navigate today’s workplace. The pattern is not only generational. Mentorship is not reassurance. It is calibration of context, perspective, and lived experience applied to your reality. The best mentoring conversations do not validate you. They reorient and broaden perspective.
Inside leadership teams. There are moments in business when you need seasoned judgment that is not on the org chart, someone to shore up a capability gap, accelerate a stalled initiative, or bring an outside line of sight to a room that has been looking at the same problem too long. As one global COO told Harvard Business Review: the smartest use of a fractional leader is not as a stopgap — it is as a strategic amplifier, a way to embed new thinking without destabilizing the core. That is not a cost play. That is a reconfiguration of what expertise looks like at the executive level.
Where AI Ends and Advisors Begin
AI is powerful. It summarizes. It analyzes. It surfaces patterns at scale. What it cannot do is tell you when the answer is not in the data but in the room or sense the quiet signals that shape real decisions.
The judgment gap does not close with better information. It closes with better counsel.
Start typing to search…