I helped grow a steel company from $750 million to $12 billion. CEOs should use AI before they grade employees on it

“Raise your hand if your executive team has mandated the use of AI.”

I was standing on stage at the annual Great Places to Work convention in Las Vegas, and every one of the 500 CHROs I was in front of put their hands up. These were all HR leaders in mid-market and enterprise companies.

“Now, keep your hands up if every single person on your executive team is using AI to make themselves more valuable.”

Almost every hand went down.

I ask this question at every speaking engagement, and I’ve seen a pattern–and a problem. Company leaders are mandating the use of AI. With that mandate, they’re expecting increases in the productivity of their workforce. But they’re not holding themselves to the same standard.

Executives often think that “leading AI” in their organizations means delegating the adoption and strategy of it to IT, a Chief AI Officer, or an outside consultant. That’s not delegation–that’s abdication.

CEOs and executive teams need to learn to use AI, and become AI-driven leaders, before they can build AI-driven organizations.

I say this as I’m starting to see companies increasingly tying AI goals and skills directly to employee raises and promotions. Should compensation be tied to AI skill? It depends on the company. If you’re an organization that typically ties performance to specific skills that will grow in value, it makes sense. If you don’t, it’s a break from precedent–but it doesn’t mean it’s the wrong move.

What does concern me is when a CEO who doesn’t use AI is the one making the call. And by “use AI,” I don’t mean to help them write better emails–I mean truly leveraging it to think strategically and make themselves more valuable. Without that experience, a leader will limit their vision for what AI can actually do for their organization. It’s tough to describe color when you’ve only seen black and white.

From search engine to thought partner

The first time I saw AI, it was December 2022. I was the Chief Growth Officer at Jindal Steel, and ChatGPT had been out for about two weeks.

When I saw AI, I saw the future.

And like most leaders, I said to myself, “I’m really busy. I’ll revisit this later.”

Like any good executive, I then thought about delegation. “I have a huge IT department,” I thought. “I’ll delegate this to them. Or maybe we will hire a consulting company to do it for us.”

Then I remembered a quote from my college mentor: “Focus on mastering the skills that are so valuable, they’ll serve you no matter where you go.” At that moment, I knew AI would be one of those skills.

That cemented my thinking. I needed to learn more about this myself, not delegate it.

I started with a simple commitment. Use AI once a day, every day. Very quickly, I realized I was using it for things that really didn’t matter, like writing better emails. Last time I checked, I never received a promotion for the quality of my emails. I had this realization that I was using AI to major in the minors. Then I started to wonder, what would it look like to major in the majors?

I started thinking about the skills I had that truly enabled me to deliver the most value. One of them was the ability to ask leaders the right strategic questions to unlock new levels of growth. In that moment, I realized I was asking AI questions like it was Google. I wondered what it would look like if I had AI ask me questions instead as a thought partner to elevate how I think and shift how I lead.

That was the game changer. I started using AI to elevate how I think, which shifted how I was leading within Jindal Steel.

A few months into my AI use, I was talking with Naveen Jindal, the Chairman of Jindal Steel. I told him how I was using AI for strategy instead of tactics, and he asked to see an example. I opened up my AI tool using the CRIT framework I’d developed, which stands for “Context,” “Role,” “Interview,” and “Task.”

CONTEXT Attached is our annual strategic plan. (Which I attached into a secure instance of ChatGPT). ROLE Your role is to act as an aggressive, growth-minded board member with deep expertise in the steel industry, vertically integrated from distribution to mining.

INTERVIEW Interview me. Ask me one question at a time, up to five questions, to understand where our plan may be insufficient to deliver results to the market.

TASK Tell me where our plan may be insufficient to deliver the expected results and the actions we can take over the next 30 days to deliver on our next earnings call.

AI started asking one sharp question at a time. Questions that only someone who knew our industry would ask. It pulled out all of the context that was in our heads. Then it laid out a real analysis of where the plan could break.

Naveen’s mind was blown. “This is incredible,” he said. I agreed and suggested we should drive it through the whole company and that he should lead it as Chairman.

His words? “I disagree. You do it!”

The problem was, I knew almost nothing about AI. But I knew the power of asking the right questions. I asked, “Who could I get into a relationship with that could enable me to drive this across our company?”

That led to a partnership with Google, where I learned a lot about the tech–and the more I learned about it, the less I realized how much the tech mattered. You can have the greatest technology in the world, but if you don’t win the hearts and minds of your people to harness the technology, you’ll never unlock the value. It became clear that AI adoption was about leadership.

Over my time at Jindal, the market cap grew from $750 million to $12 billion over a four year period. This was done with the right strategic thinking—thinking that AI directly supported throughout the company.

That is not about AI. It’s about the right strategic thinking. AI became supportive of that

I went all-in on the leadership aspect of AI, to the point where I resigned from Jindal Steel, wrote a book about what I learned, and turned it into a company.

Start with your biggest problem

One of the questions I get asked often is, “If a CEO wanted to start this week, what’s the first piece of their own work they should test with AI?”

Start by asking yourself something you’re already asking anyway: “What’s the biggest problem facing our company–the one that would unlock huge growth if you solved it?”

That question gets you focused on the 20% of work that drives 80% of your results.

Then, ask how AI can help. Write a prompt using CRIT: Describe the context in detail, tell AI what kind of expert you wish could advise you, have it interview you one question at a time, and give it a task, like, “Give me the top three high-impact, non-obvious strategies I could deploy to solve this problem.”

Here’s the biggest piece of advice: Don’t take the AI response at face value. Tell AI what you like, don’t like, and would change about the response, and keep iterating on it until you’re confident enough to act.

Inspect what you expect

When you’re using AI as a thought partner, it becomes much easier to judge whether your people are using it in a way that improves how they think and moves the business forward.

Most companies are measuring AI adoption through usage metrics (i.e. “80% of our workforce is using AI every week”). That’s a good place to start, but it doesn’t tell you whether people are actually using AI for high-value work. If you look closer, you’ll probably find that most of that usage is small tasks that don’t make anyone valuable to the company.

The better test is to inspect what you expect. When your team walks you through how they solved a problem, ask how they used AI as part of the process. The ones who have bought into AI as a thought partner will be able to tell how it sharpened their thinking, or helped them see blind spots they would have missed.

If all they can describe is time saved on emails and summarizing meeting notes, they’re complying with your adoption mandate, but haven’t really seen what AI can do for them.

You can only coach people if you’ve done the work–and the learning–yourself.

Rethink what you’re rewarding

If your company has already tied compensation or raises to AI, take another look at what you’re measuring. If it’s just usage, you’re not tracking the skill that matters. The AI skills worth rewarding are the ones that will keep growing in value: the strategic thinking, or the great line of questioning.

And if you’re a leader, be honest about whether your own pay is being held to the same standard as everyone else’s.

The next time I ask the question about whether an executive team is using AI to make themselves more valuable, I’m looking for every hand to stay up.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

This story was originally featured on Fortune.com