Every leader I speak to today is choosing between two potential versions of themselves. One is faster. The other is defensible.
I know of a leadership team that made a major market-entry decision in under an hour. A mid-market professional services firm with aspirations into a new regional vertical. It was a thirty-page AI-generated brief. The analysis was solid. Consensus arrived fast. It got approved by the CEO that afternoon.
Three months later, the decision unravelled. The faulty assumption that no one had questioned was that competition was competitive. It had been listed low on page twenty-two of 125 pages. Nobody had read page twenty-two.
That’s the velocity trap. And the AI had nothing to do with it.
The Rubber-Stamp Impulse
Boards want speed. Investors want speed. Your entire operating rhythm is built around throughput. When AI drops a credible-looking output in front of you in six minutes, the pressure to move is real. The team is watching. The window is open. The output looks credible.
So, you sign off.
I did this with a go-to-market strategy brief a few years back. My team had spent two weeks building it with AI research tools. Sixty-three slides. Market sizing, competitor mapping, channel recommendations. The most polished brief I’d seen in a long time. Everyone was aligned. I had a call to get to. I approved it in the meeting.
Eight weeks later, we discovered that a core assumption about buyer behaviour in one of the target segments was wrong. The model had drawn on industry data that was nearly two years old. Nobody had asked how recent the sources were. I hadn’t asked either.
It illustrated to me, in my very first review of anything, the difference between reviewing something and reading it.
Inside that approval somewhere, something hushes. You go from decision-maker to approver. From judgment to endorsement. What I call the rubber-stamp impulse. It isn’t laziness. When speed is the proxy for competence, and competence is exactly the thing you most need to be seen executing — decisiveness was once a sign of immovability.
But that signal is deteriorating. AI produces speed. You do not — not the same way. The exec who prides themselves on signing off with lightning speed after a model has rattled off thirty pages of analysis isn’t showing good judgment. They’re up against a tool that does not require sleep and cannot be talked down by the rest of the room.
The dangerous part is that it still appears to be leadership from the outside — for a while.
When the Team Stops Thinking
Most conversations about AI and decision quality focus on the calls that go wrong. What gets missed is quieter and more damaging — what those fast approvals tell your team about whether their own judgment matters.
If each submission receives approval within an hour, people will stop questioning their assumptions before submitting. If the answer is faster than the question, why stress-test a proposal? The quality of input declines. The room gets agreeable. And then things that are harder to bounce back from start to settle in.
There’s no announcement. No meeting where someone admits to having taken their sharp thinking off the table. It just happens. Proposals get smoother. Disagreements get softer. The choices you’re shown begin to converge towards what you’ve previously approved, since that’s the fastest path to your inbox. You’re still making decisions — you’re just making them with less friction and flimsier information.
The emotional toll of slow is profound. Pausing feels like friction. Asking the question that interrupts a meeting where everyone is saying yes feels weak. Requesting a second brief when the first one looks credible feels indecisive. So leaders absorb the cost. They perform confidence. They approve.
The emotional cost of a decision gone wrong — made in a way you can’t defend — lands later. And it lands harder than it used to.
Intentional Friction
The leaders I’ve watched get this right are doing something their peers consider inefficient. They’re intentionally creating friction in their highest-stakes decisions.
Not friction everywhere — the distinction matters. They move quickly on decisions that merit speed: tactical calls, scoped problems, issues where variables are clear and consequences are small. But for decisions that compound over time, they build in slowness deliberately and visibly.
One practice worth describing: the dissent scorecard. Before a significant decision is finalised, someone on the team is explicitly tasked with building the strongest case against it. Not to obstruct progress — to force the room to own the trade-offs before committing. It’s a rotating responsibility. Everyone learns to hold two positions at once.
The accountability test works like this: “If this fails, can I claim full responsibility without blaming the system?”
If the honest answer is “well, the model suggested it” or “the team was aligned,” more work is needed. The leader hasn’t made a decision — they’ve ratified one. That distinction sounds minor. It isn’t. When failure occurs, the team sees who steps up to the outcome and who gestures at their process.
Ownership builds trust. Speed without ownership erodes it. Slowly at first, and then fast.
Visible Leader vs. Invisible Architect
Here’s the shift most leadership conversations aren’t naming.
The visible leader model worked when authority came from presence — the executive in every meeting, running every room, whose confidence became the signal everyone else calibrated to. That model rewarded the person who performed judgment loudly and often. Decisiveness was the brand.
AI is making that model expensive.
When any meeting can be summarised before it ends, any analysis generated before the room fills, any decision benchmarked against thousands of similar decisions in seconds — the leader who holds authority through performance alone is running in a diminishing lane.
The invisible architect is what replaces it. They don’t win authority by being in every meeting. They design the decision boundary — the governance layer that determines what AI can do autonomously, what needs a human eyeball, and where real deliberation is required. They create the conditions for good judgment before the pressure comes. They’re the person who designed the feedback loop to surface failures early, and their actions prevent finger-pointing at tools when those tools fall short.
This transition requires a different relationship with visibility — a quieter ego. In organizations where the loudest decision-maker still wins, becoming an invisible architect can feel like regression. But what it looks like from the outside, over time, is a leader whose team consistently makes better decisions, including when the leader isn’t in the room. That’s the signal that’s starting to matter to the boards and investors who’ve seen the pattern: the teams that fall apart without their leader versus the ones that keep running well.
The Edge That Compounds
Global C-suite surveys found that more than half of large organizations had integrated AI into at least one business-unit-level function by 2024. Fewer than one in five had an established accountability framework to ensure those decisions were owned. The divide between adoption and governance isn’t bridged. It’s widening.
Your judgment under pressure doesn’t dilute by repetition. It compounds.
Each time you reject the velocity trap, draw the decision boundary, use the dissent scorecard, and take the accountability test, you’re building an asset your peers don’t have. While they’re hastily approving things and calling it leadership, a clean quarter will eventually produce an arduous year.
The outputs AI generates are improving. You can learn to slow down when it matters. Your real authority lies in the gap between what AI can generate and what only you — with your judgment and credibility — can back. It doesn’t display on a dashboard. It may never come up in a quarterly review. But your team feels it in how secure they are when they bring you the hard calls. And eventually, it enters the record — testable, not just well-announced.
The leaders who matter most over the next decade won’t be the fastest. They’ll be the ones whose teams can answer yes to this: Did someone who understood the trade-offs make this call?
What’s Already Ending
The era of conspicuous authority isn’t ending because leaders matter less. It’s ending because the performance of authority — the fast track to approval, the firm hand in the room, the decisive voice on a call — is no longer the signal it was.
Speed is available to everyone. Judgment under pressure is not. That asymmetry is the opportunity — but only if you’re willing to be honest about the moments your instincts have pulled you toward praise and away from responsibility.
Here’s the question worth sitting with this week.
When was the last time you rejected an AI-generated recommendation on instinct, before anyone else in the room agreed? Or asked the question that killed the momentum of a unanimous decision?
If you can’t remember, that might be the most useful thing this piece gave you.
I write about the decisions leaders avoid having. If this landed differently than you expected, or if you’re sitting with a version of this in your own organisation right now, I’d like to hear it. Drop it in the comments, or send me a DM.