Over the past couple of years, a lot of companies looked at what AI could do and made a simple calculation: If a machine can handle the work, why keep paying a person to do it? So they let go of staff, pointed to the technology, and told everyone how much leaner and faster they were about to become.
Then something funny happened. A good number of those same companies started hiring those same people back.
It didn’t get a lot of fanfare, for obvious reasons. Nobody wants to hold a press conference to announce they were wrong. But this pattern is showing up all over the place, and there’s a real lesson in it for anyone who manages people.
The Same Story, Over and Over
Take Ford. They leaned on automated systems to catch quality problems on their vehicles, the kind of thing a sharp engineer notices almost by instinct. The automation couldn’t quite do it. So Ford brought back hundreds of experienced engineers to fill the gap, and wouldn’t you know it, their quality scores climbed right back up. CNBC reported the same reversal playing out at company after company.
Or take Klarna, the online payments company. They replaced a big chunk of their customer service team with a chatbot and bragged about it. The bot was fine for the easy stuff. But when a customer showed up frustrated, confused, or dealing with a real mess of a problem, it wasn’t enough. Satisfaction dropped. Before long, the company was rehiring, with its CEO admitting out loud that customers need to know there’s always a real person they can reach.
You could line up more examples, but they all pretty much follow the same pattern. The AI handled the routine part just fine. What it couldn’t handle was judgment. Context. Reading a room. Knowing when the rulebook doesn’t apply. Remembering that the client in question has been around for nine years and deserves a little grace. That’s the stuff people carry around in their heads, and it turns out it’s a lot harder to replace than anybody figured.
None of this means AI is useless. Far from it! Used well, it takes the grunt work off a team’s plate and frees people up for the things only they can do. The mistake wasn’t using the tool but using it for everything.
An Expensive Way to Learn
Here’s the part that stings for the companies that got it wrong. Hiring people back is expensive. There are the costs to recruit them and get them up to speed again, and they often come back at a higher rate than before, because now they know exactly how much they’re needed. That’s a costly way to learn a lesson that could have been skipped.
What the Smart Ones Figured Out
The lesson in all of this isn’t complicated. Before a job gets automated away, it’s worth looking hard at what that person actually does all day. Not the tasks on paper, the real work. The judgment calls. The relationships. The quiet fixes nobody sees. More often than not, that’s the part holding everything together, and it’s exactly the part AI can’t touch. There’s more in our past article on how recognition fits into an AI-driven workplace if you want to dig into it.
The smartest companies right now aren’t the ones racing to replace their people. They’re the ones handing their people better tools and getting out of the way, and finding small ways to show them that they’re valued, whether that’s a word of thanks or a proper recognition award. Keep the humans and give them AI to make the boring parts faster. This combination actually works. And the rest are learning the hard way, one rehire at a time.