Two Good Reasons for a Layoff, and Three Bad Ones

One of my clients laid off several people because of AI.

I want to start there because I think it was a legitimate reason.

The company had people reading receipts, invoices, and expense reports and manually entering that information into a system. It was repetitive work, and AI had gotten very good at doing it.

They automated the process and measured a vast improvement in efficiency. One person could handle exceptions and check the quality of where AI was uncertain.

Work that had supported several positions could now be done with fewer people.

That’s a real business change. And I don’t think it would have been responsible for him to keep paying people to do work the company no longer needed.

The first question I ask in situations like this is whether those employees can be retrained or moved into other roles. I’ve talked before about the importance of reskilling people as technology changes.

Sometimes there are good opportunities to do that.

Sometimes there aren’t.

Technology used to be expensive to build, and relatively inexpensive labor was often the better return. AI has changed that calculation for certain kinds of work. Tasks that once required entire teams can now be automated relatively cheaply.

I don’t think CEOs should pretend otherwise.

I do think they need to be honest about what actually changed.

The Two Good Reasons

When I’m working with a CEO and CFO on a restructuring, I want to understand three things: the budget, the results we’re producing, and the strategy.

If those aren’t aligned, something has to change.

A layoff happens for one of two reasons. We can’t afford all the positions, or our strategy changed and the positions aren’t needed.

Maybe the company grew too quickly and can’t support its current cost structure. Maybe the market shifted and the strategy needs to change. Maybe we found a better way to produce the same outcome and no longer need certain activities.

My client’s was the second kind. The work went away.

At CDNOW, when I was asked to improve efficiency, my first step was to stop doing certain things.

That was important.

We looked at the work itself. Which activities were contributing to the business? Which weren’t? What could we stop doing? Cut whole departments.

There were people attached to the work we eliminated. Good people, in many cases. That made the decisions difficult.

It didn’t make the work necessary again.

Those are the two reasons I accept. Here are three I don’t.

AI Is a Very Convenient Explanation

Through September 2026, Challenger, Gray & Christmas counted 120,136 job cuts attributed to AI, roughly 21 percent of the cuts announced that year.

That’s a remarkable number.

It’s also worth understanding what it measures.

Challenger tracks the reasons employers give for layoffs. When a company says AI was responsible, that gets counted as an AI-related cut.

It doesn’t establish that AI actually eliminated the work.

There’s already a term for what happens when companies use AI to explain cuts they might have made anyway: AI washing. Even Sam Altman has publicly raised concerns about it.

I’m not inside those companies. I can’t tell you which announcements are legitimate and which are convenient explanations.

I can tell you I’ve seen plenty of businesses hire too aggressively, struggle to manage the organization they built, and then find themselves under pressure to improve profitability.

Those problems existed long before ChatGPT or Claude.

And blaming AI can sound a lot better to investors than admitting the business grew in ways that leadership didn’t know how to manage.

I’m sure there are companies making real productivity gains and adjusting staffing accordingly. My client is one example.

I’m equally sure AI is sometimes getting credit for fixing problems it didn’t create.

The distinction matters because the underlying management problem doesn’t disappear when you give it a new name.

The technology may explain why the company needs different skills or fewer positions. It doesn’t automatically tell you which employees should leave, whether someone could be retrained, or whether the business has handled the process fairly.

Those decisions still belong to management.

And if you’re blaming AI, be ready to say which activities were automated, how you measured it, and why those people couldn’t move to other work.

That’s a management decision. It deserves more scrutiny than a sentence in a press release.

Cutting Headcount Isn’t the Same as Cutting Work

The second bad reason is a number.

Some companies decide they need to reduce payroll by a certain amount, eliminate positions, and leave the rest of the organization to figure out how to get the same work done.

That’s the opposite of how we handled it at CDNOW. We didn’t just decide how many people we wanted to cut and then ask the remaining employees to somehow absorb everything.

Maybe that’s unavoidable in an immediate financial crisis. As a long-term operating strategy, it doesn’t work particularly well.

You still have to decide what the company is going to stop doing.

Otherwise, you’ve made the organization smaller without necessarily making it more efficient.

A Layoff Isn’t a Substitute for Performance Management

The third happens on a smaller scale.

A company has an employee who’s been underperforming.

Everybody knows.

Their manager has avoided the difficult conversations. Maybe they’ve offered vague feedback. Maybe they’ve convinced themselves things will improve. Maybe they’re waiting for the person to figure it out on their own.

Then the company announces a restructuring. Suddenly that position is eliminated.

Convenient.

And the employee usually knows. They ask, “Are you going to replace my role?” You say no. Then the same job reappears under a different title.

That’s not a layoff. That’s a firing with extra steps.

I understand the temptation. Managing performance takes time. You have to set expectations, give people useful feedback, document what happened, and make difficult decisions when they don’t improve.

If you’ve avoided that effort for a year or two, a layoff can look like an easier way out.

I discourage it.

What I object to is using a layoff as cover for a performance decision you were unwilling to make honestly.

Those are different things.

If the company genuinely needs fewer positions, you should be able to explain the business rationale and how the affected positions or employees were selected. If performance ratings are part of your criteria, they’d better be thorough and accurate. I didn’t use them. Performance is subjective, and a layoff has to be objective to be fair.

If the problem is that one person isn’t doing the job, that’s a performance management issue. You should have been addressing it long before a layoff gave you an opportunity to avoid the conversation.

I’ve written about why regular one-on-ones matter so much. This is one of the reasons.

A good one-on-one gives you the opportunity to say the difficult thing while it’s still small. It gives the employee a chance to respond, improve, ask for help, or understand that the role may not be working out.

Waiting until things have become intolerable doesn’t make you a kinder manager. It just makes the eventual decision harder on everybody.

The CEO Still Owns the Decision

I don’t expect CEOs to avoid every painful decision. Sometimes keeping an organization viable requires eliminating jobs. I do expect them to own the decisions they make.

My client automated a manual process, measured a substantial productivity improvement, and eliminated positions because the work had changed.

I can understand that.

If another CEO tells me AI forced a layoff, I have a few questions.

Show me what changed.
Show me what work went away.
And show me why this was the right decision for the business.

Because if the answer is really that you over-hired, mismanaged performance, or never figured out how to run the company at its current size, I’d rather hear that.

AI didn’t make those decisions.
You did.