How CEOs Should Think About AI and the Workforce

Most CEOs I speak with have already decided on AI and are under pressure to deploy it fast. They’ve approved a budget, named someone to lead it, and asked for a roadmap. The targets tend to be vague: increase efficiency, cut costs, etc. And when it comes to AI and automation, those goals typically get interpreted as reducing headcount by both leadership and the workforce. As a default activity, reducing headcount as part of AI activity is running ahead of the evidence. But how you think about AI and the workforce shapes what you get out of it.

Goldman Sachs estimates AI has trimmed U.S. payroll growth by 16,000 jobs a month over the past year, and that figure includes slowed hiring as well as outright cuts. Interestingly, 55% of businesses that have laid off employees due to AI implementations have admitted it was the wrong staffing decision.

The way to avoid joining that 55% is to start with the work before the headcount, keep reading to see how.

AI is changing the workforce, displacing some roles while creating others. The World Economic Forum projects 92 million roles will be displaced globally by 2030 and 170 million created. That’s a net gain of 78 million, with roughly 22% of all jobs churning in the process.

That churn is the real change, and looking at headcount alone hides it. A team can hold steady in size while the roles inside it turn over completely.

For CEOs, that means AI’s impact on the workforce should be looked at over time rather than through one big restructuring exercise. Understand how the work is changing first, and the decisions about roles and headcount become much easier to make.

AI changes jobs from the inside, reshaping the skills a role needs even when the title stays the same. The WEF projects that 39% of the skills people use in their jobs will change or become outdated between now and 2030. Roles can expand into more nuanced or specialized work as AI takes on the manual, repetitive tasks those roles used to contain.

Underneath that shift, individual roles move in three directions at once, and it helps to name them plainly:

This unevenness is why a blanket headcount cut tends to hit the wrong roles, trimming where work is only shifting and sparing where it is genuinely disappearing.

The mistake I see leaders make is reading “net gain” as reassurance. A net gain hides a hard truth for individual companies: the jobs lost and the jobs created don’t sit in the same place, on the same team, or in the same skill set. Managing that gap inside your own business is the work.

AI’s workforce impact touches cost structure, capability, and culture at the same time, which is why it can’t sit with one function. IT owns the tools. HR owns the training. Neither owns the decision about what your workforce should look like in three years, that belongs to the CEO.

However, there are measurement challenges to understand. McKinsey’s research found that C-suite leaders estimate only 4% of employees use generative AI for a meaningful share of their daily work. Compared with 13% of employees who report doing so, roughly three times the leadership estimate. Essentially, this difference in opinion people closest to the work are redesigning it, but leadership can’t see it happening, which makes it hard to turn into strategy.

The same research found that 92% of companies plan to increase AI investment, yet only 1% consider themselves mature in how they use it. Money is flowing. Clear direction is what’s missing.

Most companies decide where AI fits after they have already set their hiring plan. It works better the other way around. Look at what AI can do for the business first, then decide who you need to hire, who you can train, and which jobs need rethinking.

This is where the CEO and HR have to work together:

In practice, three activities can achieve this:

None of this is possible without a clear picture of the skills your people already have and the ones you’ll need next. That is where workforce planning and talent intelligence pay off. With that view, you can build a workforce that fits where the business is going, instead of scrambling to catch up every quarter.

AI keeps changing what your people need to be good at, so the workforce you build for today will not be the one the business needs in two years. Ultimately, CEOs need to work with HR to make sure ongoing training programs align with what the business will need one, two, and three years out.

That means outlining which strengths the business needs no matter how the technology turns out. Judgment, communication, and the ability to learn quickly will always be valuable in evolving markets and business environments.

The stakes here are competitive, not just operational. The WEF found that 85% of employers plan to prioritize workforce upskilling, and 63% say a shortage of the right skills is the single biggest thing holding back their plans to change the business. In other words, most of your competitors already see their people as the bottleneck and you’ll need to keep up.

When an AI-driven cut goes wrong, it tends to fail in the same three ways:

This is why the redeploy-or-discard call matters more than the size of the cut. The WEF found that 41% of employers expect to reduce their workforce as AI automates tasks, while nearly half plan to move affected staff into other parts of the business instead. The second group keeps the human context that turns an AI tool into a business result. The first group often ends up buying it back.

Thinking clearly about AI and the workforce comes down to one act of leadership that can’t be delegated: deciding which human capabilities to protect while everything around them shifts.

So here’s the one step I’d start with this quarter. Pick the three business functions most exposed to AI in your business, and for each one ask what human judgment is vital to keep it compliant, effective, and functioning cross-collaboratively with other business functions. The answer will tell you more about your workforce strategy than any tool demo.