Every HR department is scrambling to figure out how to evaluate employees in an AI world. They're building competency frameworks. They're creating certification programs. They're measuring prompt quality and output volume.
They're overcomplicating it.
After 20 years leading technology teams, I can tell you the only metric that matters: curiosity.
What this means for the CIO, CTO, and CISO
Your AI competency assessments are measuring the wrong things. Certifications, prompt libraries, and output metrics miss the fundamental differentiator.
The employees who will thrive aren't the ones who use AI the most. They're the ones who understand it most deeply. Usage is easy. Comprehension is rare.
Your competitive advantage in AI isn't technology. It's the curious people who know how to partner with it. You can buy the same tools as everyone else. You can't buy curiosity.
The inside perspective
When I built teams, the employees who thrived weren't the ones with the best credentials or the most experience. They were the ones who asked "why" and "what if" constantly. They were relentlessly curious.
I hired plenty of people with impressive resumes who turned out to be mediocre performers. They knew what they knew, and that was the end of it. When something new emerged, they waited to be trained. When something broke, they waited to be told how to fix it. They were competent within their boundaries and useless outside them.
The curious ones were different. They'd dig into problems nobody asked them to solve. They'd experiment on their own time. They'd come to me with ideas that started with "I was wondering why" and ended with "so I tried this and here's what I found."
Those people were worth ten of the credentialed-but-incurious. They're the ones who built things that mattered. They're the ones who figured out how to make technology actually work instead of just following the manual.
The outside observation
Now I watch companies try to measure AI competency with frameworks designed for a different era.
They're counting certifications. They're measuring prompt volume. They're tracking output quantity. They're building elaborate rubrics that evaluate everything except the thing that matters.
The result is predictable: organizations full of people who can use AI tools but can't make them work. Who can write prompts but can't solve problems. Who can generate output but can't generate value.
Meanwhile, the curious employees, the ones who actually understand what's happening under the hood, are being evaluated by the same metrics as everyone else. Their deep comprehension counts the same as someone else's superficial usage.
The measurement systems are optimizing for the wrong thing. And you get what you measure.
The uncomfortable truth
Machines aren't curious.
AI is a wonderful, amazing, brilliant machine. It knows the corpus of human knowledge in theory. It can process information faster than any human. It can generate output at scale that would take humans years.
But it isn't curious. It can't ask "what if we tried something completely different?" It can't wonder why something works the way it does. It can't pursue a question simply because the question is interesting.
Humans can do all of those things. The best humans are infinitely curious. They can't help but ask questions, explore possibilities, and dig deeper than they need to.
That's the differentiator. That's what makes humans valuable in an AI world. Not the ability to prompt. Not the ability to use tools. Not the ability to generate output. The ability to be curious, to ask questions that machines would never think to ask. AI can answer questions. Only humans can ask the right ones.
The productivity paradox
AI is the biggest productivity enhancement in my 38-year career. Nothing comes close.
But you don't get that out of the box.
I've watched employees treat generative AI as deceptively simple. They type prompts. They get output. They assume that's all there is to it. The output is mediocre, but they don't know enough to recognize it's mediocre. They blame the tool when the results disappoint.
Those employees are making themselves replaceable. If all you can do is type prompts and accept output, you're competing with everyone else who can type prompts and accept output. That's a race to the bottom.
The employees who get exceptional results are the ones who invested time in understanding. They learned how the models work. They experimented with different approaches. They built context over time. They treated AI as a partnership to be developed, not a tool to be used.
The difference between my AI output quality and someone else's isn't in the prompt. It's in how much I've invested in getting the AI to understand the problem at its root. When I ask a question, the AI has the right foundation to provide the right answer. It doesn't provide slop. It provides quality output.
That investment is curiosity in action. Curious people can't help but dig deeper. Incurious people stop at the surface.
The complexity hiding in plain sight
There is no piece of AI that is not far more complex than any solution that came before it.
The problem with AI is it makes things seem easy. It's the easiest thing to interface with. It's the easiest thing to get output from. Type a sentence, get a response. What could be simpler?
Everything about that ease is a trap.
If you want quality, if you really want to solve problems, if you really want to not make mistakes, all of that ease hides enormous complexity. The people who treat AI as simple will get simple results. The people who recognize the complexity and invest in understanding it will get results that actually matter.
This is the same conversation we had about data quality and small language models. If all you want is the easy answer, I'm sorry. This is too complex. You're going to be disappointed.
Smart companies are going to show they understand that complexity. They're going to demonstrate they can navigate it. And they're going to win because their competitors treated complexity as someone else's problem.
What leaders must do
Identify the curious people in your organization and enable them to work with AI.
This isn't complicated. You already know who they are. They're the ones who ask questions in meetings that make everyone think. They're the ones who experiment without being told to. They're the ones who come back with "I tried something and here's what I learned."
Give them skunkworks projects. See what they come up with.
But here's the hard part. You have to be ready for what comes next.
You have to be ready to listen. Curious people will come back with insights that challenge your assumptions. If you're not prepared to hear it, don't ask for it.
You have to be ready to change. The whole point of curiosity-driven exploration is to find better ways. If you're not prepared to change based on what they find, you're wasting everyone's time.
You have to be ready to adopt. Work with them knowing that change is coming. Go into it knowing that change will have risk. And go into it knowing you're going to adopt when you come out.
Start by scoping the skunkworks project in a small way. What's something you wish you could do that you couldn't before? Something that makes a difference, something you would actually adopt if you had a solution for it? That's your starting point. Give it to your curious people. Get out of their way.
How to tell genuine curiosity from surface-level usage
Volume of usage is not curiosity. Depth of understanding is. The curious employees you want to enable share a pattern.
- They ask "why does this work?" not just "how do I use it?"
- They experiment without being assigned to.
- They come back with "I tried something and here's what I found."
- They question whether the output is actually good.
- They dig into failures to understand root cause.
- They connect AI capabilities to unsolved problems.
- They get frustrated by shallow answers.
The surface-level users look busy but stay shallow.
- They follow tutorials step by step without deviation.
- They use AI frequently but don't understand it.
- They accept output without questioning quality.
- They wait to be trained on new capabilities.
- They blame the tool when results disappoint.
- They ask "what prompt should I use?" instead of understanding why.
- They're satisfied with "it works" as an explanation.
What I'd tell my former self
If I had known then what I know now:
I would hire for curiosity over credentials every time. Credentials tell you what someone learned once. Curiosity tells you what they'll learn forever.
I would create safe spaces for experimentation. Curious people need room to fail. If failure is punished, curiosity is punished.
I would measure questions asked, not just answers delivered. The quality of someone's questions reveals the depth of their thinking.
I would stop assuming training creates capability. Training transfers knowledge. Curiosity creates understanding. They're not the same thing.
The industrial revolution lesson
This tool is coming regardless. You don't get a vote on whether AI changes your industry. The only question is whether you're ready.
It's like walking into the Industrial Revolution, seeing a sewing machine, and going "nope, I'm a fan of my needle and thread."
I don't care if you're a fan of the old ways. This is the way we're doing things now. Figure out how to use the sewing machine or find another place to work.
There are thousands, tens of thousands, of people unemployed right now who have this attitude that AI is replacing them. And maybe it is replacing some of them. But it's not replacing everyone. It's replacing the people who refuse to adapt.
We have to change that attitude. Your attitude has to be: I do things AI cannot do, like being curious, because humans are the only things that can be curious, and AI does things I both can't and don't want to do. That's the partnership. That's how humans and AI work together. That's how you stay valuable.
The employee wake-up call
If you're an employee reading this, let me be direct.
The employees who treat generative AI as deceptively simple, who don't invest time in mastering it, who just prompt and pray, are making themselves replaceable.
It's not because AI is taking their jobs. It's because they're competing on the wrong dimension. If the only thing you bring is the ability to use a tool that everyone else can also use, you're a commodity. Commodities get replaced by cheaper commodities.
The employees who will matter are the ones who understand what AI can and can't do. Who know when to trust it and when to verify. Who can identify problems that AI can solve and problems that require human judgment. Who are curious enough to keep learning as the technology evolves.
Those employees aren't replaceable. They're force multipliers. They make everyone around them more effective. They're the ones who figure out how to actually get value from AI instead of just generating output.
Which one are you?
The 2026 prediction
Curiosity becomes the differentiator in 2026.
The hype will cool. The easy wins will be exhausted. The companies that just deployed AI because it was trendy will face the reality that deployment isn't the same as value.
The organizations that thrive will be the ones that identified their curious people early. That enabled them to explore. That listened to what they found. That actually changed based on what they learned.
The organizations that struggle will be the ones that measured the wrong things. That counted certifications instead of comprehension. That rewarded usage instead of understanding. That optimized for metrics that had nothing to do with value.
Companies that enable curious workers with AI will beat companies that try to replace workers with AI. Every time. Without exception.
The curious will inherit the AI earth.