I spent a good part of this week in conversations with some senior leaders, and I noticed something none of them quite said out loud.

They believe that their teams are producing better work than ever. The decks are sharper; the analyses faster. The first drafts land almost finished. By every visible measure, the organization looks more capable than it did eighteen months ago. 

And yet, underneath the polish, I could sense that there was something hiding in plain sight, something that has the potential to lead to catastrophic results if not addressed soon. When I reflected upon this, I realized that the risk I saw was my sense that none of these senior leaders could now tell how much of that capability is their team’s vs how much of that is from the AI tool that they’ve been using. 

And I think it’s a bigger strategic problem than most leadership teams have understood or have given serious thought about. 

The mirage

Researchers at MIT Sloan Management Review gave named this phenomenon: the capability mirage.

Here’s what is happening. AI-assisted (and in most cases, AI generated) output looks professional regardless of the skill that produced it. The junior analyst and the seasoned one now hand you work that reads about the same. 

The output converges. The underlying capability does not.

You are suddenly in a position where you are unable to judge how much of the work was informed by the expertise in your team vs how much of it came from the tool. You will find it difficult to judge whether your team members are getting better at understanding the business, gleaning insights, learning and understanding more deeply about the business or are just getting better at using the AI tool. 

So, the signal you’ve always relied on, the quality of the work in front of you, stops telling you what it used to. You can no longer read capability off the output your team produced. And because everything looks good, you will end up not looking for the change in capability. 

That’s the mirage. Not that the work is bad. That the work looks uniformly excellent while the judgment beneath it quietly thins out and you have no dashboard that shows you this is happening.

The part nobody wants to say

Now here is where it gets uncomfortable, because the instinct in every organization is the same: we need to upskill the workforce.The data says the problem is somewhere else.

A recent study by HFS Research and EY of over 300 large-company executives found that 77% of leaders say they know what role AI will play in their future, but only 13% have a working human-and-AI operating model running. This gap alone is sobering.

But the finding that stopped me was this one: the people least prepared for the shift weren’t on the front line. They were the senior leaders and the middle managers. The steering layer. The people sponsoring the transformation. The very people who are the making the most consequential decisions are the least equipped to not only understand and leverage the technology being deployed across the organization but are also ill-equipped to judge if the very people they are leading are getting better and more effective at their work or are getting better at using the tool they are using. 

Sit with that for a moment. 

The mirage isn’t just below you. It’s around you. And it may be loudest at the top.

Why this is a strategy problem, not a training problem

It would be easy to file this under “learning and development.” I don’t think that’s where it belongs.

When access to AI is universal, the technology stops being the differentiator. Everyone has access to the same models. Everyone gets the same polished first draft.

What’s left as a durable advantage is judgment, understanding the context, ability to glean insights. The ability to look at a good-looking answer and know it’s wrong. The ability to read a room, to sense which risk is worth taking and which is a bet you’ll quietly regret, to know what your business needs when the tool confidently recommends something adjacent. Differentiate between noise and signal. Evolve our sense making ability. 

The tool comes after the sense making is done, not before. 

That ability of sense making is the compounding asset. It’s the thing competitors can’t buy off the shelf. And it’s precisely the thing the mirage erodes without ever showing up on a report — at every level, including among the leaders. 

To be clear about what I’m not saying

I’m not arguing for less AI. I use it every day, and I’ll keep using it. It frees me to spend my attention where it matters.

I’m also not romanticizing the way we used to work, when capability was easier to see because everything took longer.

What I’m arguing for is narrower and, I think, more important. There’s a difference between AI that multiplies your people’s judgment and AI that quietly substitutes it. 

In the short term, the two look identical. Both produce better-looking work, faster. They only diverge over time, when one organization has deepened its judgment and the other has outsourced it without noticing. It shows up in moments of real crisis. 

The organization that has used the technology to multiply and improve the sense making abilities of their teams will thrive and the other will struggle to come out of the crisis, as the sense making abilities of their employees would have atrophied over time. 

By the time that divergence is visible in the results, it’s very expensive to reverse.

A simple test

So, here’s what I’d invite you to do this week.

Look at the last three pieces of excellent-looking work that crossed your desk. For each one, ask: Do I know how capable the person behind this is? How much of this work is based on the competence of the person who created this and how much of it is dependent on the tool that was used? 

Then turn the spotlight on yourself and your leadership team. On the decisions that mattered most this quarter, the ones that needed real judgment, how many did you fully think through before reaching for the tool?

If those questions are harder to answer than they should be, that’s not a failing. That’s the mirage doing exactly what it does. The leaders who will pull ahead in the next few years won’t be the ones whose organizations look the smartest and the one’s using the latest and greatest technology available. 

They’ll be the ones who can still tell the difference between looking capable and being capable. They will be the ones who will continue to invest in the people as much as they invest in the tools, deliberately. 

I’d genuinely like to hear how you’re seeing this in your own organization. Are you finding ways to measure real capability under the output? Or does the mirage have you too?

Being aware of this mirage and becoming deliberate in our ability to recognize the risk helps us become a leader worth following.