The New Leadership Equation: Why AI Makes Judgment More Valuable Than Speed

Deepali Kaushik, executive at NBCUniversal, is an Amazon bestselling author and writes about technology, leadership and transformation.

gettyI’ve had to make decisions where the fastest solution wasn’t the right one. A release might be technically ready, but another round of testing could uncover a problem that would be far more expensive after launch. A process might appear efficient, but a missing safeguard could create a privacy or security risk. A technology migration might meet its timeline while creating operational dependencies that hadn’t been considered.

AI is changing the scale of this challenge because it can analyze enormous amounts of information, identify patterns and recommend actions in seconds. But it creates a leadership question I believe we aren’t asking often enough: When technology can move faster than human beings, what should leaders still be responsible for?

In my work, I’ve learned not to treat a dashboard, report or automated result as the decision itself. The data can tell me that something changed. It can’t always tell me why it changed or what the consequences of acting on it will be.

The Lesson: Before accepting an AI-generated recommendation, ask: What information is this based on? What might be missing? What would make this recommendation wrong?​

When something affects a production environment, customer experience, security posture or privacy obligation, I don’t ask only, “Can we do this now?” I ask, “What happens if we’re wrong?”

Sometimes, the right answer is to accelerate. Sometimes, it’s to run another test, involve another expert or add another control. And sometimes, the right leadership decision is simply to say no.

The Lesson: Establish explicit pause points for decisions involving material operational, security, privacy or customer risk.​

Technical solutions aren’t always the hardest part—it’s deciding what the technology should do in the first place. Ownership can’t disappear simply because a process becomes automated.

Consider an app store age-restriction decision. Deciding whether the product should be available to users aged 13 and older or restricted to those aged 18 and older is a different kind of challenge.

A 13+ approach could expand access to a younger audience, but it could also introduce significant implementation complexity. In the U.S., children’s privacy requirements and other applicable state laws can create overlapping obligations, depending on the product, its users and how their data is collected and used. Meeting those obligations can affect product design, privacy controls, operational processes and ongoing compliance.

An 18+ threshold could simplify some aspects of serving younger users, but it could also exclude a generation of potential customers who might otherwise grow up using the product.

Neither option is just a coding decision. AI can analyze the trade-offs, but leaders must balance legal risks, costs, customer acquisition and long-term growth and own the consequences of their decisions.

The Lesson: For every AI-enabled process, identify three things before deployment: who owns the outcome, what decisions AI is permitted to make and when a human must intervene. That’s accountability by design.​

Technology leaders naturally measure efficiency, time saved, cost reduced, processes automated and capacity increased. When evaluating a technology change, I also ask what happens to the people around the process. Does the employee gain capability or simply lose a task? Does the customer experience actually improve? Does the organization reduce one type of risk while creating another?

A system can be technically successful and still produce a poor leadership outcome.

The Lesson: Add a human-impact question to technology reviews: Who’s affected by this change, and how might their experience change? It takes very little time to ask, but it can change the quality of the decision significantly.​

Organizations are understandably focused on AI adoption. Leaders are asking where AI can save time, reduce costs and increase productivity.

But capability alone isn’t a strategy. Instead, ask: Where should we use AI, and where should we deliberately keep a human in the loop?

That distinction requires experience. A leader who understands only AI capability may automate aggressively. A leader who understands only risk may avoid innovation. The better approach is to evaluate both.

The Lesson: Three questions help move AI enthusiasm to responsible adoption. What value will AI create? What risk could it introduce? What decision should remain human?​​

The more I work with technology, operations, quality, cybersecurity, privacy and AI, the more I’m convinced that leadership in this era requires three things: capability, judgment and responsibility. Capability defines what technology can do, judgment guides what we should do and responsibility determines what we own.

AI will continue to make organizations faster. It will make information easier to access and execution easier to automate. But those advances don’t make leadership less important. They make the quality of leadership more visible.

When a machine can generate 10 possible answers in seconds, the leader’s value isn’t simply producing an 11th answer. It’s knowing which questions need to be asked, which answers deserve to be challenged, what risks haven’t yet been considered and when decisions should remain human.

That’s why I believe the leadership advantage of the AI era won’t be speed alone. It will be the judgment to know when speed creates value and when wisdom requires a pause.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?