The Restructuring Is Real. Whether You Should Worry Depends on One Thing

By mid-2026, the euphemisms had mostly given up. Companies used to blame "macro headwinds" or "reprioritization." Now they just say AI. Oracle cut 21,000 jobs. Amazon cut 16,000. Meta cut 8,000. Block halved its headcount, with its CEO telling shareholders that "a significantly smaller team" using AI tools "can do more and do it better."

AI isn't necessarily the sole reason these jobs are disappearing — companies have plenty of other levers to justify a cut. But it's changing the calculation behind restructuring: in some functions, a team that once needed ten people to cover a workload may now plausibly cover it with six, and cover more ground while doing it — potentially more output, faster turnaround, fewer hours logged, similar quality on the routine slice of the work. Atlassian's CEO was unusually candid about this when 1,600 roles went: "it would be disingenuous to pretend AI doesn't change the mix of skills we need or the number of roles required." That's the calculus now sitting on every CFO's desk — not "can AI do this job," but "how few people do we need to get the same output, and can we defend that headcount to the board." Man-hours are being repriced in real time, and the repricing doesn't ask permission.

Should You Actually Be Worried?

Not equally. Layoffs are never perfectly discriminating — plenty of skilled, ambitious people get caught in cost cuts and reorgs that have nothing to do with them personally. But one group is particularly exposed: the ones who've gotten comfortable. Comfortable in a role that hasn't changed in three years. Comfortable with a skill set that was cutting-edge once and has been coasting since. Comfortable assuming their function is too specialised, too relationship-driven, or too senior to be touched. Every one of those assumptions is being tested right now, and "I've always done it this way" is not a defence any board is interested in hearing.

The uncomfortable truth is that the people most exposed aren't necessarily the least skilled — they're the least current. Skills don't decay because they stop being true. They decay because the direction the company is heading in moves past them. If your skill set isn't actively tracking where your organisation is heading, you're not standing still, you're falling behind, just slowly enough not to notice until the org chart changes.

What I've Done About It

I started out in enterprise risk management, covering information security and operational risk for systems — that's where I learned how risk actually gets identified, owned, and escalated inside an organisation, which turned out to matter more than I expected later. When the opportunity came, I jumped into cybersecurity, on the simple bet that demand in the field was only going one direction. That's where I built on those fundamentals: security hygiene, risk ownership, and the practical levers that actually improve an organisation's security posture.

From there I moved into second-line risk management at a private financial institution, overseeing tech, cybersecurity, and — increasingly — AI risk. That's about as close to the heart of the current conversation as a risk function gets right now.

Here's the part I keep having to relearn: none of that is enough on its own. Hard skills got me into the room. They don't decide whether anyone listens to me once I'm there. Knowing when to push hard on a colleague and when to stay friendly, reading a difficult conversation and knowing what to say and what to hold back, presenting myself so a technical point actually lands — that's the layer hard skills can't substitute for, and it's the layer I'm still actively working on. This blog is part of that work. I started it as a way to document what I'm learning as I go, and, honestly, it doubles as evidence — for myself as much as anyone reading my resume — that the learning is ongoing, not something I finished once and stopped.

Standing Out From Your Peers

Staying current with the direction of travel gets you to "not at risk." It doesn't get you to "indispensable." The officers who'll actually stand out are the ones who pair technical currency with something AI is structurally bad at: soft skills.

This is the trade-off worth understanding clearly. Some technical tasks are increasingly being commoditised by AI — not because technical depth stops mattering, but because AI narrows the gap between a strong technical operator and an average one, at least on the routine slice of the work. Soft skills don't compress the same way. AI is genuinely getting good at drafting communications, role-playing difficult conversations, even coaching a presentation. But it doesn't carry the accountability or the organisational trust that comes from actually being in the room. It can help you prepare for a hard stakeholder conversation. It can't have the conversation for you, and it can't be the one people decide to trust when the model's output needs a human sign-off.

The differentiators worth building are unglamorous but durable: presentation skills that hold a room, stakeholder management that treats a difficult conversation as a negotiation rather than a briefing, influencing skills that get buy-in without needing authority, and the ability to explain a technical risk in one sentence a non-technical stakeholder will actually remember. None of that shows up on a certification. All of it shows up in whether people trust your judgement when the model's output needs a human sign-off.

The honest starting point is auditing your own profile — not generically, but specifically against where your organisation is heading. Most people can name their technical gaps quickly. Fewer can honestly say whether they're a strong influencer or just a fluent talker. Know which one you actually are before someone else finds out for you.

There's one more thing worth naming, and it's the piece that took me longest to see in my own path: moving through risk, into cybersecurity, into technology and AI risk didn't just teach me each discipline separately — it taught me how they connect. Someone who understands enough security to know what's actually at risk, enough of the business to know what's worth protecting, and enough people to get something done about it is harder to replace than someone who's simply excellent at one of those things alone. The future probably doesn't belong to the most technical person in the room. It belongs to whoever can hold the technical, the risk, and the human parts of the conversation at the same time.

Takeaways

1. Get honest about whether your current skill set is tracking where your organisation is heading, or just where it used to be

2. Invest deliberately in the soft skills that don't compress under AI — stakeholder management, plain-language translation of technical risk, and the judgement calls a model can't yet own

3. Audit your own strengths and weaknesses honestly. Comfort in a lane you've mastered is not the same as safety in a lane that still needs you