Why HR Must Shift from Skills Management to Capability Design

The World Economic Forum predicts 39% of global workers’ core skills will change by 2030. Additional research by Stanford University’s Digital Economy Lab found a roughly 13% relative decline in employment among early career workers in the occupations most exposed to artificial intelligence (AI), while employment for more experienced workers held steady, meaning the occupation stayed the same, but who got to enter it did not.

What this data points to is something a skills framework has nowhere to record: not whether a skill exists, but whether a person still owns the work.

Where the Model Breaks

Years ago, my marketing team used to put graduates through a small test. Arrange some data in a pivot table. We were not checking whether they knew pivot tables. We were checking whether they could quickly figure out something they had never done before. The pivot table was never the point. The point was the thing underneath it, the confidence to step into an unfamiliar tool and make it work.

That test is worth noting because, for a decade, HR has treated skills as an inventory problem: map them, catalog them, find the gaps and then buy the training to close them. It was the right discipline for a stable world — one where the skills a business needed changed slowly enough that keeping the list current was a reasonable ambition.

Now, AI is breaking this model.

Watch what happened to that aforementioned test. The pivot table went first — a task AI now finishes before the graduate has even opened the file. Then, last year’s version of the skill went, the integration between two tools no one bothered to list anymore. Next, it will be the confidence to open a coding assistant and let it carry real work.

Letting the pivot table go was the right call. Some work should leave human hands, and pretending otherwise is nostalgia. What you want to catch is the different thing: the work that slips out while the label stays the same. A skills inventory records what can be done. It can’t tell you when a capability is being hollowed out behind an unchanged label.

This is what I call a “custody problem.”

How to Make the Shift: The Custody Audit

The shift doesn’t require a wholesale redesign. It starts with a diagnostic, or custody audit, you can run against your existing framework. Most audits tell you what your people can do. This one is about what they still do themselves, and what has drifted over to the machine without anyone deciding it should. Before the next skills-gap exercise, work through your critical roles and ask these three questions:

• Who does this work now — a person or a tool? Mark what your people still hold and what has passed to a tool while the job description stayed the same. The gaps that matter are the capabilities you are handing over without deciding to.
• If the tool does a task, do we still need a human who could do it? Some of the overage is fine to release. In some cases, though, you need a person kept close enough to recover when the tool fails.
• Is AI taking the rungs judgment grows on? The junior tasks that AI absorbs first are often where senior judgment is built. Take away the bottom rungs and you can unwittingly protect this year’s output while dismantling next year’s expertise.

Finally, act on what the audit shows. Name the few capabilities you will protect and keep the list short enough that the board can hold on to it. Build roles that keep people close enough to the work to step in. Reward the capability, not only the finished output.

Taking the Decision to the Board

Skills management was built to answer what your people can do. The question now is what skills are still theirs, and how many of those skills are you handing over to AI without ever deciding to. That isn’t a training question, and it’s bigger than any learning and development budget. It’s a decision about where an organization’s judgment will sit once AI does far more of the work, and that decision should be made at board level.

Keep counting, and you may end up managing an inventory of work the machine already does. The harder job is to work out what human capabilities are worth protecting, while there is still a choice to make. That is the real decision HR should bring to the board.

Editor’s Note: Additional Content

For more information and resources related to this article, see the pages below, which offer quick access to all WorldatWork content on these topics:

• Career Growth
• Organizational Performance
• People Analytics