A few months ago, I watched a strategy head for a global consumer brand drop a 40-page, AI-generated competitive analysis into a leadership meeting. It was synthesized overnight and more thorough than her team could have produced in two weeks. The room went quiet, though not because anyone disagreed with the findings. The team members had spent their careers being rewarded for finding the right answer, and now it had arrived before anyone sat down. All that was left to do was react.
One senior director said what everyone was thinking: “So … do we just agree with this?”
That moment is playing out in organizations everywhere. A November 2025 Wharton survey of more than 800 senior leaders found that 82 percent use generative AI weekly and 43 percent already worry it’s eroding their teams’ skills as adoption climbs. That’s the tension underneath the excitement: Teams are leaning on AI faster than they’re building the judgment to use it well.
In optimizing for execution, most organizations trained the curiosity out of their people. The technology simply made it visible, but curiosity is the one thing AI can’t replicate.
Why expertise became the enemy
Experience, track record and pattern recognition form what I call the River of Thinking. The longer someone has worked in an industry, the faster that current flows, matching new challenges to old solutions. It’s efficient, often correct and, by definition, incapable of producing anything genuinely different.
AI learned from the same river. It was trained on published thinking, historical data and accumulated best practice. Ask it to analyze a market, and it hands the river back faster and more confidently than any analyst could, which is a ceiling as much as a capability. Another Wharton study found that while AI improves individual ideas, teams that rely on it generate ideas strikingly similar to one another, shrinking the variety that drives breakthrough innovation. As Christian Terwiesch, co-director of Wharton’s Mack Institute, noted, “If you rely on ChatGPT as your only creative advisor, you’ll soon run out of ideas, because they’re too similar to each other.”
How to lead in the space AI creates
Rebuilding the conditions for genuine human thinking isn’t a cultural initiative or an off-site. It’s a small set of decisions a CEO makes and models personally, every day:
1. Surface the river before you act on the analysis. When AI-generated work lands in a room, the instinct is to evaluate it: Do we agree or disagree? That’s the wrong first move. Before any analysis drives a decision, name the assumptions underneath it explicitly. Ask your team: “What would have to be true for this to be wrong? What did this model almost certainly not account for? What does our experience tell us that the data can’t?” AI can tell you what the river looks like. Only your team can ask whether you should be in a different river entirely.
2. Model intellectual curiosity. The most powerful signal a leader sends isn’t in their strategy deck. It’s what they do when they don’t have an answer. Leaders who respond to uncertainty by projecting confidence—even performed confidence—teach their teams to do the same, and over time, organizations lose the ability to sit in a question long enough to find something genuinely new. I didn’t understand this until a colleague pointed it out to me directly, years into leading creative teams at Disney. I’d developed a habit of tilting my head slightly whenever I liked an idea. I never said anything. But my team had mapped it, and they’d started pitching toward the tilt instead of the problem we were trying to solve. The moment I heard that, I stopped taking notes in meetings, just to remove every visible tell I had. Ideas got stranger and better almost immediately.
3. Design the boundary between AI and human thinking deliberately. Most teams use AI to produce outputs they then react to. Flip the sequence. Use AI to generate the constraints, the baseline and the data, and then create explicit space where the team’s job is to question, reframe and imagine beyond it. I picked up a version of this from a conversation with Richard Branson years ago, and I still use it with teams today. Before anyone looks at an idea—AI-generated or otherwise—we agree on what “good” actually means for that specific decision, and write it down where everyone can see it. Only once that’s settled do we let the ideas, or the AI’s output, into the room. It sounds obvious, but almost no team does it in this order. Without it, whoever talks first or loudest usually wins the room. With it, an idea has to earn its place against a standard nobody could quietly change after seeing what came out of the model.
The silence was the signal
I’ve thought about that room a lot since I left it. Forty pages of flawless analysis, and silence. The silence was a symptom of an organization that had optimized relentlessly for answers and never invested in questions.
AI didn’t create that problem, but it’s making it urgent, and it’s a leadership problem before it’s a technology one. Winning this era means using what AI gives back—time, bandwidth, cognitive space—to rebuild the human capabilities no model can replicate.
The question isn’t whether your organization can use AI. By now, most can. The question a CEO should be asking the board is what the team does when the analysis is done and the room goes quiet.
