The New Leadership Advantage

The advantage moves from answers to judgment.

Monday morning. A CEO is asked to approve a discount to save a large deal, settle a dispute between two leaders and decide whether to expand into a new market. The dashboard is current. The AI assistant has produced recommendations. The calendar is full.

Information is not the problem. The CEO must still decide which question matters, whose evidence is reliable, which risks are acceptable and what the organisation is willing to stand for. An answer can be immediate while the decision remains difficult.

A decision is a choice. Judgment is the quality of reasoning that makes the choice defensible.

AI can help gather evidence, generate alternatives and test scenarios. But a fluent recommendation is not proof that its assumptions are sound—or that its objective is the right one. A system asked to increase conversions may recommend an action that damages contribution margin. It has answered the question it was given; the leader must examine whether that was the right question.

The six inputs to judgment are context, evidence, mental models, principles, past decisions and reflection. They are complementary. Evidence without context can answer the wrong problem. Experience without reflection can repeat an old mistake. Principles without attention to consequences can leave important trade-offs unexplored.

The point is not to slow every choice. Routine, reversible decisions can often be delegated or automated within clear boundaries. Reserve deeper reasoning for choices with substantial stakes, uncertain causes or consequences that are hard to undo. In those choices, accountability remains with the people and institutions making them.

The discount recommendation.

Start by separating the symptoms. Did conversions fall because price became less competitive, because the team attracted lower-quality leads, or because proposals arrived too late? Compare similar deals within segments. Review lost-deal reasons and the time between enquiry and proposal. An overall conversion rate can conceal very different problems.

At a 10% discount, contribution per sale falls from ₹30 to ₹20. The team needs 150 sales instead of 100—a 50% increase—just to preserve total contribution. That does not prove discounting is wrong. It makes the condition for success explicit, so the leader can test whether the expected volume increase is credible.

A defensible next step might be a short diagnosis followed by a limited experiment. Name an owner, a decision date and the evidence that would justify expanding, changing or ending the test. Judge the action against its objective and guardrails; do not celebrate conversion gains while ignoring their cost.

Take one decision off autopilot.

Choose a consequential decision you face this week. It might be a pricing change, a hire, an investment in a new capability or a personal commitment. Spend ten minutes on the note below. You do not need to finish it to finish reading this chapter.

1. Which decision am I postponing—and what makes it difficult?
2. What assumption am I treating as a fact?
3. What evidence or argument would genuinely change my mind?

Think of your last major success. Was the reasoning sound, or did circumstances happen to favour you? Ask the same about a failure. A good outcome can hide weak reasoning; an unfavourable outcome can follow a well-reasoned choice.

Record your reasoning before the outcome is known. Return on the review date and compare what you expected with what happened. Look for mistaken assumptions, missing evidence and influences outside your control. The aim is to improve the next decision, not to construct a flattering explanation of the last one.

The leadership advantage is not having the best answer on demand. It is knowing how to question, test and act on an answer.