Ask a leadership team whether they plan to augment their people with AI or replace them, and you will usually get a confident answer. Ask them to show you where that decision was made, and the room gets quiet.

That silence is the problem. In most companies, nobody chose. The strategy emerged from a series of small calls: a role left unfilled here, a contractor not renewed there, a team told to “do more with the same.” Two years later, someone looks up and discovers the company has a workforce strategy. Nobody wrote it.

Drift Is a Decision Too

Every workforce move you make now is quietly one of two bets. Augment: keep the people, raise what they can produce. Replace: shrink headcount, hold output steady. Both are legitimate, and neither is automatically right. But treating them as the same thing, or as something that will sort itself out, is how companies land somewhere they would never have approved in a board meeting.

The short-term math on replacement is easy to love. If one AI-assisted person can do the work of three, you save two salaries, and the margin improvement shows up on a slide. Augmentation is harder to put in a spreadsheet cell, which is exactly why it gets undervalued.

What the Spreadsheet Doesn’t Show

Spreadsheets are great at counting what left and terrible at counting what stayed. Augmenting keeps things that never get a line item. Someone remembers why the process works the way it does. A customer trusts one particular person, and that trust took years to build. Someone catches the AI’s error before it ships.

The asset I think leaders underweight most is spare capacity. A company that cut its way to AI productivity has nothing to deploy when the market suddenly opens up. A company that augmented has a team ready to say yes.

Then there’s the operational side. AI works best when paired with people who can correct it and override it. Remove them and you remove the safety net.

The same tool that makes an augmented team faster can produce errors at production scale in a team hollowed out around it. The tool didn’t change. The configuration did.

So ask a different question than “what’s cheapest this year?” Ask what mix of people and AI holds up across the next decade. Companies thinking in quarters tend to replace, and companies thinking in cycles tend to augment. I can’t prove it yet, but my bet is that the second group is more likely to own the next cycle.

To be fair, I’m not arguing that replacement is never right. Some roles will shrink, and pretending otherwise is its own kind of dishonesty. My argument is narrower: if a role is going to shrink, that should be a choice you can explain and defend, with real transition support for the person in it. It is much easier to stand behind a path you picked than one you drifted into.

Make the Choice for One Role This Week

You don’t need a task force. You need one role and an honest hour. Here is the version you can run today:

  • Pick one role where AI is changing the work. Not the easiest, not the hardest. One that matters.
  • Split the work. What share is routine and pattern-based? What share requires judgment, relationship, creativity, or ethics? Be skeptical of “mostly judgment.” It rarely is.
  • Write down the path and the reason. Augment or replace, and why. One paragraph is enough.
  • Attach an action. If augmenting, what training, tooling, and time are you committing this quarter? If replacing, what support does the person in that role get?

The last step keeps this from turning into a philosophy exercise. A choice that produces no action is just a preference. You will probably choose differently for different roles, and that’s fine. Matching answers across roles matter less than having a reason for each one.

If you do this and discover you were already replacing without having decided to, that is useful information, not a failure. It means you now get to decide on purpose.

Where to Find the Rest

The book spends more time on the people side of this shift, including the fears employees carry into it. They are not all the same fear, and they don’t all have the same answer. That treatment lives in AI Strategy for Business Leaders, and you can find the free tools at my companion resources. If you want to know where my perspective comes from, my background has the short version, and my books covers the rest of the shelf.

Pick one role. Make the call. Write it down.