Six months ago, a company laid off half its customer support team and told investors AI would handle the rest. Last month, they posted new job listings for… customer support.

Sound familiar? It should. It’s becoming one of the most predictable stories in tech.

The pattern ✨

It plays out almost the same way every time:

1. A company announces AI will take over a function — support, content, coding, ops.
2. Staff are laid off. Investors cheer. Headlines call it “the future of work.”
3. Six to twelve months pass.
4. AI turns out to handle roughly 60% of the job well. The other 40% — judgment calls, edge cases, the messy human stuff — doesn’t go away.
5. The company quietly rehires. Sometimes the same people. Sometimes new hires doing the exact same job with a different title.

No press release for that part.

Why this keeps happening?🤔

AI is genuinely good at the repeatable 60%. It’s the last 40% that keeps tripping companies up — the judgment calls, the exceptions, the moments a customer needs an actual human to say “I understand, let me fix this.”

Leadership teams are optimising for a headline (“we’re an AI-first company now”) faster than they’re validating whether the technology can actually own the full job. That gap between announcement and reality is where the boomerang happens.

There’s also a simpler explanation some are starting to say out loud: AI has become a convenient story for cuts that were coming anyway. As one prominent investor put it recently, most large companies are overstaffed — and “AI” is a much better headline than “we hired too aggressively during the boom.”

Why this matters if you’re leading a team 🫡

If you’re weighing an “AI will replace this function” decision right now, the rehire boomerang is worth sitting with before you act:

– **Test before you cut.** If AI is only proven to handle 60% of a role, you haven’t automated the role — you’ve automated part of it. Know exactly which part before you make headcount decisions.
– **Rehiring costs more than not firing.** Severance, lost institutional knowledge, recruiting, onboarding, and the morale hit of “we’re back” all add up to more than the salaries you thought you were saving.
– **The trust cost doesn’t rehire itself.** Customers, employees, and investors remember the announcement. They don’t always see — or believe — the quiet correction.

AI or humans? 🤖

AI absolutely can replace parts of jobs, and pretending otherwise is its own kind of denial. But “parts of jobs” is not the same as “jobs.” Companies that treat the two as identical are the ones ending up back where they started — just with a bigger bill and a harder conversation with the team they let go.

The smartest move isn’t “AI or humans.” It’s knowing exactly where that 60% line sits in your business before you bet headcount on it.

Source: Forbes