Why Contextual Judgment Is A Critical Leadership Skill AI Cannot Replace

Artificial intelligence is becoming extraordinarily good at giving us the "right" answers. In fact, MIT warned this week that frontier models can deliver "correct" responses to most undergraduate college assignments.

AIs can summarize a market, model a scenario, interrogate a strategy, generate a campaign, identify patterns in customer behavior, and now even make serious progress on frontier mathematics.

That is extraordinary. It is also where one of the most important leadership challenges of the AI age begins.

AI can help us understand what is happening, what has happened before, and what may happen next. Leadership requires something more demanding: deciding what should happen.

This distinction is classic yet also an incredibly contemporary insight for every leader working with AI. It sits at the center of my work as an AI keynote speaker.

Why can’t AI replace leadership judgment? AI can analyze what was, identify what is, and predict what could happen. Leadership requires deciding what ought to happen in a unique context, integrating evidence with purpose, values, relationships, intuition, lived experience, and responsibility for the consequences.

AI and the Gap Between “Is” and “Ought”

Almost 300 years ago, Scottish philosopher David Hume highlighted a deceptively simple problem.

We can collect facts about what is true in the world, but facts alone do not logically tell us what ought to be done. To move from an “is” to an “ought,” we also need values, priorities, purposes, assumptions, and judgments about what matters.

The "is-ought problem" is intensely practical for every leader and decision-maker.

Data can tell you that a division is underperforming. AI can benchmark the numbers, model redundancies, and estimate the savings from removing 15 percent of the workforce. None of those calculations can settle whether that is the wisest move for this company, with these people, this strategy, these customers, this culture, the specific history, and alternative, possible, preferable futures.

AI can estimate which marketing campaign concepts resemble successful past campaigns. It cannot finally determine what your brand should stand for, which convention is worth breaking, or which strange idea could become the breakthrough everybody else failed to see.

The same applies to product innovation, culture change, hiring, mergers, transformation, and strategy. The leadership challenge begins where the facts stop being sufficient.

AI Is Brilliant at Prediction. Leadership Requires Judgment.

At the core of large language models is probabilistic prediction. Given a sequence of tokens, a language model predicts what is likely to come next.

Modern AI systems are far more sophisticated than that sentence can capture. They can use tools, search, generate reasoning traces, test possibilities, and verify outputs. But probabilistic next-token prediction remains foundational to how LLMs work.

This makes AI extraordinarily powerful at discovering and recombining patterns across enormous quantities of accumulated human knowledge.

This is why I increasingly think of AI as Alternative Intelligence: not a replacement for human intelligence, but a powerful co-creative intelligence that can challenge, extend, and accelerate our thinking while remaining guided and guard-railed by human judgment.

Even if AI becomes dramatically better at mathematics, science, coding, forecasting, and formal reasoning, leadership still contains a fundamentally different category of challenge.

Science, data, mathematics, and the smartest AIs can tell you what was, what is right now, and what could be.

Leaders must decide what is best, what is wisest, and what should be right here, right now—and make a decision that may have unintended consequences, both for the good and for the bad.

There is no theorem that proves whether you should protect a struggling but potentially brilliant employee, shut down a legacy business, challenge the CEO, invest in an unproven category, or pivot a beloved product before the market forces you to.

These contextual judgments require deep, deliberate thought—and the best data available.

The Boeing Warning: When a Proven Playbook Meets the Wrong Context

The new Netflix documentary Freefall: A Reckoning for Boeing brings this problem painfully to life. After the 1997 merger with McDonnell Douglas, former GE executive Harry Stonecipher became Boeing’s president and COO, and in 2005 Boeing appointed James McNerney, another longtime GE executive, as CEO.

Both were shaped by the Jack Welch era at GE, when its market capitalization rose from around $14 billion to more than $400 billion. This was in part driven by Welch’s famous 20/70/10 system, which ranked employees annually and routinely removed the bottom 10 percent. The model was widely admired for driving performance and cost efficiencies. The market loved it.

Now imagine a powerful AI advising leaders at Boeing at the time. Looking at the available evidence from GE, it might reasonably have concluded that this kind of management thinking works. GE had created extraordinary shareholder value, its management methods were celebrated, and its leaders were being copied across corporate America. But the data could not settle whether the same philosophy should be transplanted into an aerospace company where engineering judgment, accumulated technical knowledge, psychological safety, long-term trust, and obsessive attention to quality can literally be matters of life and death.

I am not suggesting that Welch’s management philosophy directly caused the 737 MAX disasters and Boeing's sustained challenges. Complex failures rarely have one cause. But the consequences have been devastating: two crashes in 2018 and 2019 killed 346 people, the aircraft was grounded worldwide, Boeing incurred billions of dollars in debt, the share price is half what it was, and they agreed to pay more than $2.5 billion to resolve a U.S. criminal fraud conspiracy charge.

The lesson is that AI can tell us what worked elsewhere, identify patterns, and can predict outcomes based on the past. Leaders still have to ask and answer the questions datasets cannot correctly predict: will this be valuable here, with these people, in this culture, facing these risks, at this particular moment? That is the gap between what worked and what we ought to do, and it is precisely where contextual judgment becomes indispensable.

Context Is Where Leadership Lives

Contextual judgment means interpreting information within the living reality of a specific moment.

The context includes the data, but also the people in the room, the history nobody wrote down, the trust that has been built or broken, the promises already made, the organization’s purpose, its risk appetite, the power dynamics, the timing, the emotional temperature, and the future you are attempting to create.

NIST’s work on human-AI interaction makes a closely related point. It warns that converting complex human and social phenomena into mathematical representations (which is what LLMs do) removes necessary context. Effective human-AI decision-making therefore requires attention to human dynamics, real-world impacts, values, and context-specific norms.

This is why my LEADERSHIP-AI Synthesis makes distinctions that I think will become increasingly important:

AI supplies information. Leaders create interpretation

AI suggests what is. Leaders discern what ought to be. AI evaluates general, probabilistic, and median cases. Leaders must judge the specific, unique, and idiosyncratic case in front of them.

I invite leaders to use AI aggressively where it expands intelligence, challenges assumptions, exposes blind spots, accelerates experimentation, and improves the quality of thinking. But the stronger AI becomes, the more important it is to understand exactly which part of our work we are delegating.

Why AI Advice So Often Becomes Bland

Ask an AI to analyze positioning options, pressure-test a strategy, expose weaknesses in an argument, or show you what you may have missed, and it can be exceptionally useful.

Ask it, without sufficient context, “What should we do?” and, in my experience, the answer often begins to regress toward the plausible, familiar, balanced, and defensible.

The answer can also shift significantly as prompts and contextual information change.

That makes sense. A system trained on patterns in existing human output can synthesize what has worked, what is commonly recommended, and what appears coherent.

Breakthroughs frequently require a leap beyond the bland, the average.

The campaign that changes a category may initially sound odd. The product innovation that creates a new market may violate what customers currently say they want. The intervention that unlocks a stuck executive team may depend on saying something inappropriate in almost any other room.

Context changes what is right. If AI can only tell you what was right in the past, like delivering an ideal undergraduate essay, it has value with major limits for leaders.

It is also why Adaptive Intelligence becomes increasingly important as Artificial Intelligence becomes more powerful. AI can help us find better answers to the questions we give it.

Adaptive Intelligence helps leaders recognize when reality has changed so profoundly that they need to question the assumptions, reframe the problem, and ask an entirely different question.

An AI+Leadership Judgment Protocol

A major danger is that leaders gradually outsource the very muscles required to recognize when the obvious, historically correct answer is inadequate.

When using AI to support consequential decisions, leaders need a disciplined way to benefit from machine intelligence without outsourcing human judgment. Here is my current protocol:

1. Establish what is

Use AI to build the richest possible picture of reality. Ask it to research, analyze, compare, forecast, critique, destruction-test assumptions, surface anomalies, and identify patterns. Separate what is known from what is inferred, predicted, or uncertain.

2. Understand the living context

Ask what is distinctive about this situation. What history matters? Which relationships, tensions, power dynamics, cultural norms, risks, or tacit knowledge could materially change the answer? Identify what AI cannot know because it has not lived inside this organization, market, team, or moment. Understand the data's limits and imperfections.

3. Sense Internal & External weak signals

Bring embodied and experiential intelligence into the process. Notice intuitions, tensions, inconsistencies, discomfort, excitement, and possibilities that have not yet become explicit arguments. Treat them as signals to investigate, not unquestionable truths.

4. Explore what Ought to be

Return to leadership, enterprise purpose, and values before choosing a course of action. What are we actually trying to create? What is our North Star? Which values, commitments, stakeholders, and long-term consequences should shape the decision? This is the move from AI-assisted analysis of what is toward human judgment about what ought to be.

5. Make A contextual judgment

Integrate the evidence, the specific context, the purpose, and the human signals. Decide what appears wisest here and now, rather than what is merely most statistically probable or conventionally defensible. Deliberate with others.

6. Experiment, learn, and update

In genuine complexity, certainty is usually impossible in advance. Turn the judgment into an intelligent experiment, observe what actually happens, gather new evidence, and revise quickly.

Developing these capabilities takes more than learning how to prompt an LLM. Through Switch On Leadership’s customized and experiential AI leadership development programs, we help executives develop their own LEADERSHIP-AI Synthesis, combining practical AI fluency with contextual judgment, creativity, adaptive intelligence, relational intelligence, and the transformational leadership capabilities needed to turn AI investment into real-world value.

How an AI Keynote Speaker Should Help Leaders With Judgment

This is why I believe the best AI keynote speakers should go well beyond demonstrating tools, trends, prompts, agents, and productivity hacks.

Those things matter. Organizations increasingly need AI fluency.

The deeper leadership challenge is learning to collaborate with extraordinarily powerful machines without letting our own capacity for contextual judgment, creativity, courage, and agency atrophy.

In my work as an AI keynote speaker, leadership keynote speaker, innovation keynote speaker, and transformation keynote speaker, I help leaders understand where AI can genuinely outperform us, where human judgment still matters profoundly, and how to combine the two through the LEADERSHIP-AI Synthesis.

The ambition is to help people become dramatically more capable by working with AI while strengthening the contextual judgment, imagination, relational intelligence, intuition, courage, purpose, and wisdom that let them guide it.

As AI makes intelligence abundant, the "right" answers will become cheap.

The premium will increasingly be found in asking better questions, understanding context, creating meaning, choosing what matters, making courageous judgments, and taking responsibility for what ought to happen next.

AI can help us see more clearly what is. Leadership is still the work of deciding what ought to be, and then having the courage, commitment, and perseverance to bring it into existence.

FAQs: AI & Leadership Judgment

What is contextual judgment in leadership?

Contextual judgment is the ability to interpret facts, predictions, and possibilities within the unique realities of a particular situation. It combines evidence with history, relationships, culture, purpose, values, tacit knowledge, intuition, and an understanding of what leaders are trying to create.

Why can’t AI replace human judgment?

AI can analyze enormous amounts of information, identify patterns, compare options, and predict likely outcomes. But consequential leadership decisions also require judgments about what ought to happen, which depend on purpose, values, context, relationships, risk, and consequences that data alone cannot settle.

How should leaders use AI for decision-making?

Leaders should use AI to research, analyze, forecast, compare, critique, uncover patterns, and challenge assumptions. They should then interpret that information within the living context, explore what ought to happen, incorporate weak signals and human insight, make a contextual judgment, and test it through intelligent experimentation.

What is the AI and Leadership Judgment Protocol?

Nick Jankel’s AI and Leadership Judgment Protocol has six stages: establish what is, understand the living context, sense internal and external weak signals, explore what ought to be, make a contextual judgment, and experiment, learn, and update.

What is the Leadership AI Synthesis?

The LEADERSHIP-AI Synthesis is Nick Jankel’s framework for combining Artificial or Alternative Intelligence with human leadership capabilities. AI contributes analysis, prediction, pattern recognition, and automation, while leaders contribute contextual sensemaking, purpose, creativity, relational intelligence, ethical judgment, imagination, and wisdom.

Why is contextual judgment becoming more important as AI improves?

As AI makes information, analysis, and plausible answers increasingly abundant, advantage shifts toward determining which questions matter, understanding what is distinctive about a situation, choosing between competing possibilities, and accepting responsibility for what should happen next.

Can AI make strategic leadership decisions?

AI can be an exceptionally powerful strategic partner, but leaders should be cautious about delegating final judgments to it. Strategy often involves novel circumstances, conflicting stakeholder needs, incomplete information, cultural subtleties, and decisions about preferable futures rather than simply predictions based on the past.

What should an AI keynote speaker help leaders understand?

A strong AI keynote speaker should go beyond tools, prompts, agents, and productivity. Leaders also need to understand where AI genuinely adds value, where its limitations matter, and how to combine machine intelligence with contextual judgment, creativity, adaptability, courage, relational intelligence, and transformational leadership.