A certification costs real money and usually real weekends too, so out of all of them, which one should you do next?

Something interesting is happening, and a lot of the decisions based in AI are being reversed. Understanding “why” puts you in a stronger position, whether you’re job hunting now or if you want to figure out how to be more valuable at your current position.

The pattern has been clear: a company announces AI will handle part of the work, headcount goes down, and a few months later the same work is being advertised again, sometimes under a shinier job title. This is happening mostly in operations, support and people teams, but engineering is experiencing something similar.

The clearest case of AI layoffs: operations

According to Ravio’s 2026 tech hiring data, operations hiring across European tech fell 20% in a year. They also had the worst odds of a raise, only 14% got one, and only 3.4% got a promotion. Not a great year to be in ops, on paper.

And yet, operations still has a 27% hiring rate, higher than every function except commercial. So companies are cutting and hiring for the same roles at the same time, how does that work?

As Ravio explains, routine operational tasks are getting automated or outsourced, while the more strategic side of ops is becoming more valuable. There’s a simpler explanation too, and Marie Richter, an analyst quoted in the same report, says:

Companies overestimate how much AI can replace operational and support roles. People Ops gets cut first when budgets tighten, the team quickly ends up understaffed, things start going wrong, and the roles come back. 

Other sources support Mary’s claim:

• Gartner expects that by 2027, half of the companies that blamed headcount cuts on AI will rehire for similar work, often with a different job title. 
• Robert Half found 32% of managers had cut a role citing AI, then hired someone back for it or something very close to it. 
• Orgvue reports that 55% of employers who fired people based on AI, admit they already regret those cuts.

Is quality disappearing from the workplace?

AI handles the routine volume nicely, but there’s still work that needs human judgement. 

• Ford, for example, spent 3 years bringing back around 350 experienced engineers, after its automated quality-control systems failed to deliver. Many of Ford’s most experienced engineers had already left before their knowledge could be captured by the systems meant to replace them, so the automated tools ended up reinforcing flawed assumptions instead of catching defects.As Ford’s VP of vehicle hardware engineering puts it: “AI is a fantastic tool, but it’s only as good as the information you train it on.” Those veterans ended up coming back to help us train AI and mentor younger engineers.
  
• Commonwealth Bank of Australia is the case that started this conversation, back in July 2025. It cut 45 customer service roles, saying its new AI voice bot had reduced call volumes by 2,000 a week.The Finance Sector Union disputed that arguing volumes were in fact climbing, with staff offered overtime, and team leaders pulled onto the phones, and took the bank to the workplace relations tribunal. A month later, CBA reversed the redundancies, called it an error, apologised to the staff involved, and admitted its assessment hadn’t properly considered the business needs.

In both stories, the technology didn’t fail. The business managers didn’t take into consideration the full extent of people’s jobs, just the visible, repetitive part.

The junior talent gap is already upon us

Nobody claimed AI would replace developers outright, but that senior developers plus AI wouldn’t need as many juniors. Ravio’s data shows entry-level hiring down 73% in a year, with engineering among the hardest-hit functions. 

And our own Tech Talent Trends 2026 report shows where that leads: only 12% of the tech workforce now has under 3 years of experience. When that layer disappears, so does the mentoring, the code review culture, and all the slow, unglamorous knowledge transfer that turns juniors into the seniors everyone will be competing to hire later. 

IBM looked at the same pressure and went the other way. Its internal HR assistant, AskHR, already answers 94% of employee questions without escalating to a specialist.

Instead of cutting junior hiring, IBM tripled entry-level hiring in the US for 2026. Its HR chief, Nickle LaMoreaux, reckons that the junior job of 2 or 3 years ago can mostly be done by AI now, so they rewrote the role instead of deleting it. 

A junior developer who used to spend 34 hours a week coding is now out with clients or building new things. She says: 

“If we don’t continue to invest in entry-level hires, what happens in 3–5 years? There’s no pipeline, the well simply dries up”.

The Tech Talent Trends found 74% of tech professionals already use AI coding tools, and 60% think those tools make them more productive. But there’s still 32% that aren’t sure where any of this is going.  

Meanwhile, the companies leaning hardest on AI aren’t necessarily shrinking their teams. 

Y Combinator reported that around a quarter of one recent batch had codebases roughly 95% AI-generated, and those startups kept hiring engineers. Reasoning models aren’t good at debugging, so you need people who understand the product in depth. Y Combinator partner Diana Hu made the same point:

“However much you lean on AI, the skill you can’t skip is reading code and finding bugs.”

What this means if you’re job hunting right now

• The work that needs human judgement isn’t going anywhere. Reviewing, integrating, deciding, and understanding the business is still important, hard to automate, and what companies are short of.
• The roles coming back are not identical to the ones that were cut. A lot of them now combine doing part of the job with supervising the AI doing the rest. If you’re interviewing for a role that was recently “automated,” ask directly what changed and what didn’t. It can tell you a lot about whether the company has rethought the job or just relabeled it. 
  
• If your old employer comes calling, you’re not in a weak position. They now have proof you were harder to replace than they thought. Treat it like any other offer, ask what’s different this time, weigh the risk of returning somewhere that already got this decision wrong once, and negotiate having that in mind.
  
  
• If you’re earlier in your career, make the non-automatable part of your work visible. Mentoring, context-building, catching the thing the system missed. None of that shows up in a ticket count, but it’s the part companies are discovering they cut too fast.  

AI turned out to be less of a replacement for people, and more of a way of finding out which parts of a job nobody had properly understood.