I've sat in the meeting where a role gets cut in under ten minutes.
There's a spreadsheet on the screen. Names, titles, salaries. Someone points at a row and asks, "What does this person actually do?" Nobody in the room can answer with any confidence. And somehow that silence gets treated as evidence.
I've been on the wrong side of that call. I've signed off on a cut that looked clean on paper, then spent the next two quarters finding out what that person had been holding together.
That's the lesson I keep relearning. The job is what's on the org chart. The work is what actually happens. They're rarely the same thing.
What leaders think they're cutting
When we look at a role from above, we see a title, a job description, and a cost. It feels like enough to make a decision.
Here's what actually happens. Job descriptions get written once, usually to get a requisition approved, and nobody touches them again. Meanwhile the person in the seat spends years absorbing whatever the system around them fails to handle.
They become the one who knows why that client gets invoiced differently. The one who catches the order before it ships wrong. The one three other departments quietly call when the official process doesn't work.
None of that is in the description. None of it shows up in a headcount review. All of it leaves when they do.
The bill comes later
Here's the part people miss. The cost of a bad cut almost never shows up right away.
The first month looks great. The savings are real and they're on the report. Then small things start slipping. A renewal gets missed. A handoff breaks. A good client gets a bad answer from someone who didn't know the history.
By the time the pattern is obvious, nobody connects it to a decision made a quarter earlier. It just looks like the team got worse. So we push the people who are left a little harder, and the best of them start looking around.
Most teams don't struggle after a cut because the people who stayed are lazy. They struggle because the work didn't go anywhere. Only the person did.
Watch the film first
In fifteen years of coaching football, I never cut a player off the stat sheet.
The stat sheet tells you who touched the ball. It doesn't tell you who made the block that sprung the run, or who got the defense lined up right before the snap. Some of the most valuable players I coached barely showed up in the box score. You only saw it on film.
An org chart is a stat sheet. If you want to know what a role is really worth, you have to watch the film.
What the AI numbers are telling us
This matters more now that AI is in every staffing conversation.
Through October, U.S. employers have cited AI as the reason for 121,401 job cuts this year, according to Challenger, Gray & Christmas. That's roughly one in five of all announced cuts.
Now look at what the analysts tracking those decisions are finding. Forrester reports that 55 percent of employers regret laying people off because of AI, and it forecasts that half of AI-attributed layoffs will be reversed. Gartner predicts that 30 percent of the people let go in AI-driven cuts will need to be rehired by 2029, often at a higher cost.
Those are analyst surveys and forecasts, not settled fact. But they point in the same direction, and it's the direction I'd expect.
I work with these tools every day, and they're good. They can take real tasks off real people. But a tool replaces tasks, and you can't hand it tasks you don't know exist. Automate the visible part of a job without understanding the rest, and you haven't removed a role. You've removed the person who was covering for your process.
The regret in those numbers isn't about the technology. It's about cutting the job before anyone understood the work.
What I'd do before any cut
This isn't an argument against ever cutting. Sometimes the business needs it, and pretending otherwise isn't leadership either. It's an argument for earning the decision.
-
Sit with the work, not the description. Spend real time with the person, or at least with their last few months of email, tickets, and calendar. Write down what they actually touched.
-
Ask who depends on them. Skip the manager and go to the people upstream and downstream. "What would you stop getting if this seat were empty?" is the most useful question I know.
-
Give every task a destination. For each thing the role does, name what picks it up: a person, a tool, or a deliberate decision to stop doing it. "The team will absorb it" doesn't count.
-
Find the judgment calls. Tasks transfer. Judgment built over years doesn't, at least not quickly. Know which one you're losing.
-
Price the transition honestly. If the handoff takes sixty days and you planned for five, the savings number is wrong.
The question underneath
I get the pressure. The number is due, someone above you wants an answer, and the spreadsheet is right there. Understanding the work takes time nobody feels like they have.
But the work doesn't care what the org chart says. It still has to get done by someone.
Before the next cut, it's worth asking one thing. Do I know what this person does, or do I only know what they're called?
Sources: Challenger, Gray & Christmas, October 2026 job cuts report (published November 5, 2026). Forrester, Predictions 2026: The Future Of Work, as reported by The Register (October 29, 2025). Gartner prediction on AI-driven layoffs, as reported by CX Dive (September 11, 2026).