Inside the AI boomerang: Ford's costly rehires vs. Ikea's reskilling bet

Over the past few years, many companies adopted AI and cut roles based on the assumption that AI could fully take them over. Now, some are reversing those decisions.

A growing number of companies are rehiring for roles they previously eliminated, often at a higher cost. Others, like Ikea, are taking a different approach by reskilling employees whose roles are changing.

This emerging pattern has been described as the "AI boomerang."

For HR and L&D, the question is not only which tasks AI can perform, but what happens to the people who previously performed them.

Ford's costly correction

When Ford built AI into its design and quality systems, it let go of experienced engineers whose judgement those systems still needed.

After veteran staff left, Ford's AI tools began amplifying weak designs instead of catching errors, and vehicle quality dropped.

Ford has since rehired 350 engineers, known as its "gray beard" engineers, to help rebuild the data feeding its AI systems and mentor junior staff.

Ford is not the only company reversing an AI-driven workforce decision.

Klarna replaced 700 customer service agents with an AI assistant between 2022 and 2024, then began rehiring humans in 2025 once quality dropped.

IBM has also announced plans to triple entry-level hiring this year, reversing course on roles it had marked as easy to automate.

The AI boomerang is becoming a wider pattern

Recent data shows how widespread this pattern has become:

• Almost a third of employers have eliminated a position after adopting AI and then rehired for it.
• Forrester forecasts that roughly half of AI-attributed layoffs will eventually be rehired.
• Boomerang hires return with an average 5% pay increase, compared with 2% for employees who stayed.

The pattern points to the same issue: roles can be cut based on what AI could theoretically do rather than what it has actually been tested to do.

When expertise leaves with the employee, organisations may discover that they still need it.

Why AI layoffs can turn into rehires

The problem is not simply whether AI can perform a particular task. It is whether organisations understand everything that surrounds that task.

That can create a cycle:

• AI is introduced
• The role is cut
• Human expertise leaves
• Problems appear
• The role is rehired

This is the AI boomerang.

Ikea chose reskilling instead

Ikea faced a similar starting point but chose differently.

When its customer service bot Billie launched in 2021, it could resolve less than half of incoming queries. That figure now sits at 74%.

Instead of cutting the roles Billie made redundant, Ikea decided to retrain around 8,500 call centre employees.

The training covers room-planning tools alongside judgement skills a bot cannot replicate, such as knowing which follow-up question to ask when a customer cannot explain what is wrong with their kitchen.

The reskilled team now works as design consultants and complex-query specialists.

Reskilling can create new value

Ikea's reskilling approach has produced measurable results.

Its remote sales centres are now its fastest-growing sales channel, increasing by 15–20% year over year.

Sales reached €1.25 billion last fiscal year, compared with €1.08 billion the year before.

Its in-house customer happiness score also climbed to 89%, up from 60% before Billie launched.

What HR and L&D can learn

The difference between Ford's rehires and Ikea's retention was not the technology. It was what each company chose to do with the people whose roles were changing.

For HR and L&D, the challenge is deciding:

• Which groups should be upskilled first?
• What capabilities should employees develop?
• What is a realistic first step?

‍

Two starting points are highlighted:

• Prepare managers to lead AI adoption
• Build AI capability across the wider workforce

AI transformation changes roles. Organisations need to decide what happens to the people in those roles.

The choice can be to cut and potentially rehire later, or to reskill people before their existing roles disappear.

The lesson for HR and L&D

The AI boomerang shows why workforce decisions cannot be separated from AI decisions.

The question should not only be whether AI can perform a task, but also what happens to the people whose roles are changing.

AI transformation does not have to mean replacing people. It can also mean changing what people do and developing new capabilities.

The organisations that get this right may be the ones that understand not only what AI can automate, but also what their people can do next.

‍

‍