Stories about large companies cutting IT staff while hiring people with stronger AI skills hit a nerve, and not just for the people directly affected. Whether it’s a carmaker, a bank or a retailer, the pattern is similar: shrink some roles, grow others. If you work in technology and you’ve read a headline like that, you’re probably wondering what it means for you. This piece isn’t about any one company’s decisions. It’s about the broader shift and what practical steps make sense.
Why companies restructure around skills
Technology teams in big organizations often carry layers of roles built for an earlier era: maintaining legacy systems, managing on-premise infrastructure, running manual testing, handling routine support tickets. As automation and cloud services take over parts of that work, the mix of needed skills changes. At the same time, leaders feel pressure to show progress on AI, so they seek people who can build, evaluate and deploy it.
Cost matters too. Restructuring is sometimes a way to cut expenses with a forward-looking label attached. Workers are right to be skeptical when a layoff announcement is wrapped in language about transformation. The honest reading is usually a mix of real strategy and financial pressure.
Which tasks are really exposed
Automation tends to hit tasks, not entire jobs, but when a role is mostly one task, it’s vulnerable. Work at higher risk includes:
- Repetitive ticket handling and basic helpdesk responses.
- Routine scripting and report generation.
- Simple code translation and boilerplate writing.
- Basic data entry and cleanup.
- First-pass documentation.
Work that holds up better involves judgment, context and accountability: designing systems, understanding business needs, security decisions, vendor negotiation, incident response and managing risk.
What ‘AI skills’ usually means
The phrase is vague, and employers often use it loosely. In practice, it can cover several different things:
- Using AI tools well. Prompting, checking output, and weaving assistants into daily work. This is the most widely needed skill.
- Building with AI. Connecting models to company data, creating retrieval systems, writing evaluation tests.
- Data skills. Cleaning, structuring and governing data, since models are only as good as what they’re given.
- Machine learning engineering. Training and tuning models, a deeper specialty.
- Governance and risk. Privacy, security, compliance and responsible-use policies.
You don’t need all five. Pick the one closest to your current work and build from there.
A realistic upskilling plan
Start by applying AI to your existing job. If you write SQL, try using an assistant to draft and review queries, then note where it’s wrong. If you manage infrastructure, use it for scripts and documentation. This gives you real examples to discuss in interviews.
Next, build a small project. It could be a tool that summarizes internal documents, a script that classifies support tickets or an evaluation harness that tests an AI feature. Show your work in a portfolio or on a code hosting site if your employer permits.
Then fill in fundamentals. Understanding how models behave, where they fail, how to measure quality and how to protect data counts for more than memorizing the latest product names. Courses help, but a project teaches faster.
Protect yourself in the meantime
- Keep your resume current and document your accomplishments with concrete results.
- Maintain your professional network outside your employer.
- Build savings if you can, to give yourself options.
- Learn your employer’s policies on severance and benefits before you need them.
- Stay cautious about feeding confidential data into public AI tools.
What employers should consider
Replacing staff with new hires isn’t the only path. Retraining existing employees often costs less and preserves institutional knowledge, which is hard to hire. People who understand the company’s systems, customers and quirks are valuable when paired with new skills. Employers who invest in internal training also tend to see better morale and retention, which matters when a company is trying to make a big technical shift.
Looking ahead
No one knows exactly how this will unfold, and be wary of anyone offering certainty. Some roles will shrink, others will appear, and many will change quietly. The most durable position is to be someone who understands both the work and the tools, who can judge whether an AI output is right, and who takes responsibility for the result. That combination is difficult to automate, and employers keep asking for it, whatever the job title says.
Featured image: "Kept In" (Jessie Wilcox Smith).jpg by Jessie Willcox Smith, Public domain, via Wikimedia Commons.
