The Cost of Being Wrong

A few years ago I watched a leadership team make a decision in about ninety seconds.

The question on the table was whether to pause hiring for a team they expected automation to absorb within a year or two. Someone laid out the case. Heads nodded. The logic was clean, the slide was tidy, and the room moved on to the next item before the coffee had cooled.

It looked like decisiveness.

I've come to think it was something closer to the opposite.

What nobody asked, in that ninety seconds or after, was the only question that actually mattered: What if we're wrong?

Not "are we wrong" — you can't know that yet.

But if the timeline slips, or arrives faster than anyone in the room expects, what does that error cost us, and who feels it first?

That question rarely gets asked. It sounds like a hedge: the cautious item you tack onto the end so you can say you considered the risks.

It isn't a hedge.

It's the question that quietly governs all the others.

And while AI happens to make the stakes unusually visible, the same question sits beneath most consequential leadership decisions.

---

Every Position Is a Forecast

Strip away the confidence and strategy language, and every stance a leadership team takes on AI and human work is a prediction about timing.

How fast this moves.

How far it reaches.

Which roles change, and when.

We dress these predictions up as plans, but underneath, they're forecasts -- guesses about a future nobody has seen.

And a forecast can be wrong in two directions, not one.

You can move too fast — restructure around a transition that stalls, let go of people whose work turns out to still be needed, hollow out capability you can't easily rebuild.

Or you can move too slow — hold a structure long past its usefulness, absorb cost your competitors shed, and make the hard changes anyway, later, under greater pressure and with less room to be humane about them.

Most teams plan for only one of these.

They choose a central estimate — their best guess at the pace — and build around it as though it were a fact.

The estimate might even be good.

---

The Two Errors Don't Cost the Same

The two ways of being wrong aren’t symmetric. But the asymmetry isn’t the obvious one.

The tidy version goes like this:

Move too slow and the cost lands on the organization — margin, market position, a few quarters of catching up.

Move too fast and the cost lands on people — the role cut a year early, the team rebuilt around a tool that wasn't ready — costs that don't sit on a balance sheet but in someone's living room.

That's only half right.

Move too slow for long enough and the cost reaches people too. It just arrives later, and often all at once.

The company that holds its structure past its usefulness doesn't absorb the gap indefinitely. It absorbs it until it can't, then corrects hard — deeper cuts, less warning, less room to be humane about any of it.

A few issues back, in When Innovation Outpaces Governance, I called this the danger of reassurance by analogy — the quiet confidence that because past disruptions eventually sorted themselves out, this one will too. That comfort is the too-slow error in its most respectable clothes.

So the real asymmetry isn't organization versus people. Both errors can reach people. The asymmetry is in three things the tidy version hides:

Timing. Move too fast and the harm is immediate and certain. Move too slow and it's deferred and probabilistic — which is why it's easier to dismiss.

Reversibility. A too-slow position can often still be corrected before it reaches anyone. A role cut too early often cannot be uncut.

Attributability. Harm from moving too fast is traceable to a decision you made. Harm from moving too slow tends to arrive wearing the costume of necessity — something the market did to you, not something you chose — which makes it the error you can disown. Hold onto that one. It comes back.

When you only model your central guess, you miss all of this.

You're holding a timeline when what requires judgment is the shape of the downside on either side of it, how reversible each error is, and who is standing underneath when it falls.

---

Who Bears It First

That last question is the uncomfortable one, so it's the one worth sitting with.

If your timing is off, the people who often feel it first are those with the least influence over the decision and the least capacity to absorb its consequences.

That's not a moral failing of any particular leader. It's the gravity of how many organizational decisions distribute their effects.

And this is where attributability returns.

A leadership team that has genuinely reckoned with this can say, out loud, who that person is in their case.

A team that hasn't will reach for the passive voice:

"The technology made this necessary."

"The market forced our hand."

They let the decision dissolve into something that merely happened.

---

The Two-Direction Test

None of this works as an abstraction. The way to make it real is to stop saving the question for a planning offsite six months out, and run it against a decision in front of you.

Here's the move, in four questions:

Run those four on a live decision and the question stops being philosophy. So do the disagreements it surfaces — and the disagreements are the point. If the room agrees instantly, you've asked it too vaguely to bite.

---

---

Deciding Deliberately

I'm not arguing for caution. Moving too slow is a real error with real victims, and "let's wait and see" is not the wise, responsible position it often pretends to be.

I'm arguing for something narrower and harder: that you price both errors before you act, and name who absorbs each one.

Leadership isn’t eliminating uncertainty. It’s choosing which uncertainty you’re willing to own.

None of it slows the technology down. That was never on offer. What it changes is whether the decision is yours or the pressure’s — whether you chose the bet you're holding, or simply inherited it from the tidiest slide in the room.