When data feels decisive, leaders stop asking harder questions

Risk has never been more measured. Or more misunderstood.

Across finance, policy, and enterprise leadership, the prevailing wisdom is clear: more data leads to better decisions. Better models mean fewer surprises. AI promises to remove uncertainty altogether. In boardrooms and risk committees, confidence now often arrives wrapped in dashboards and forecasts.

But that confidence is fragile.

The last few years—geopolitical shocks, supply-chain fractures, banking stress, and whiplash economic cycles—have exposed a gap between what models predict and how reality behaves. The issue isn’t that leaders lack information. It’s that many have outsourced judgment to systems designed for cleaner worlds than the one we’re operating in now.

As Evan Sekeris, Senior Vice President and Senior Economist at the Bank Policy Institute, puts it: “There is no smoke without a fire. AI is the real deal—but it doesn’t replace human understanding.”

At ZRG, we see the same pattern across talent and risk decisions. The firms pulling ahead aren’t rejecting technology. They’re re-anchoring it to human insight. They know the future of risk leadership isn’t about predicting every outcome. It’s about understanding what models can’t tell you—and acting anyway.

AI didn’t change risk. It exposed how leaders misunderstand it.

AI is accelerating a trend decades in the making: more data, more modeling, more speed. But speed without context creates a new vulnerability. Models are exceptional at identifying patterns. They are far less capable of interpreting meaning when the environment shifts.

Sekeris frames this clearly: today’s risks are structural, not cyclical. Geopolitical instability isn’t noise—it reflects a fundamental change in global power. AI itself isn’t just another tool; it’s as economically transformative as the industrial revolution.

In these moments, backward-looking data becomes less reliable. What worked before may actively mislead now. Leaders who treat AI outputs as answers instead of inputs mistake precision for certainty.

The most valuable risk leaders understand why risk exists—not just how to price it

One of the quiet failures in modern risk management is over-specialization. Technical skill is abundant. Conceptual understanding is rarer.

Sekeris argues that economics—not just math or modeling—creates durable risk leaders because it teaches what risk represents, why it’s priced the way it is, and how incentives shape behavior. Without that foundation, organizations become excellent at optimizing within assumptions they never question.

We see this in executive hiring, too. Boards often look for leaders who’ve “seen the model.” The better question is whether they’ve challenged it. The strongest leaders don’t just manage variance. They interrogate the system producing it.

Technology scales efficiency. Judgment determines outcomes.

AI is an efficiency engine. It surfaces insights faster, flags anomalies earlier, and expands what leaders can see. But it cannot interpret incomplete information, navigate moral tradeoffs, or weigh second-order consequences.

Sekeris is explicit: technology will not replace human understanding or the ability to react under uncertainty. In high-stakes environments, judgment—not computation—is still the decisive advantage.

ZRG’s work with boards and investors reinforces this daily. The leaders who outperform don’t ask, “What does the data say?” They ask, “What might the data be missing—and who is equipped to see it?”

Where the counterargument holds—and where it breaks

Yes, some environments benefit from full automation. Highly stable, rules-based systems reward speed and consistency over interpretation. In those cases, human intervention can introduce bias or delay.

But those environments are shrinking.

As markets fragment and shocks compound, the cost of misinterpreting context grows. AI excels inside boundaries. Leadership begins where boundaries blur. ZRG’s view holds because risk today is less about deviation and more about direction.

The takeaway for leaders

The tension we started with is unresolved for a reason. Risk cannot be engineered away. It must be understood.matt

Leaders should stop asking how advanced their models are and start asking who is accountable for interpreting them. Build teams that blend technical fluency with economic intuition. Hire leaders who are comfortable making decisions when the data is incomplete and the stakes are real.

Because in the new risk environment, confidence doesn’t come from certainty. It comes from judgment.

‍