Ask a leader if they're using AI well, and you'll mostly hear yes. Ask them how they know, and the answer gets thinner.
AI's effect on your thinking is mostly invisible to you. You see the summary, the draft, the time back. What's harder to see is how the output is quietly shaping what you decide, and what you stop questioning along the way.
Watch how leaders interact with AI over a few months and three patterns show up. You'll probably recognize at least one in yourself, or in the leaders you coach.
Competing.
You see how quickly the model summarizes a forty-page document, and something in you wants your team to match it. The grinding eats the time you used to spend on judgment and conversation, which is the work that's actually growing. Collaborative work has climbed more than 50% in two decades and now takes 80% of the average knowledge worker's week (HBR). You're racing on AI's track and losing ground on yours.
Deferring.
The output sounds confident, your time is short, you skim and ship. In a recent Wharton study, four out of five knowledge workers followed an incorrect AI answer even when they could've caught it. The decision goes wrong in exactly the place no one was watching, because it sounded like it didn't need watching.
Ignoring.
Maybe you're skeptical, cautious, or just too busy. Whatever the reason, you're now firmly in the minority. Three out of four knowledge workers are already using AI at work, and 78% of them are bringing their own tools when no one supplies them (Microsoft Work Trend Index). The gap between how your team thinks about AI and how you do is becoming the one that defines what they bring to you, and what they don't.
Each pattern carries a cost, both for you and for the people you're trying to lead. Underneath all three sits the same dynamic: AI mirrors how you think. You don't get smarter when you use it. You get faster, and you amplify what you were already doing.
The Whole Brain® Thinking framework names four thinking styles tied to how the brain prefers to work: analytical, practical, relational, experimental. We all lean toward some and underuse others, a pattern the HBDI® assessment makes visible. Here's the catch. We verify hardest in the styles we enjoy and skim through the ones we'd rather not look at. AI then hands those underused styles back to us with full confidence, and we accept the output where we should be questioning it most.
What does that mean for your leadership?
Four practices that move you past competing, deferring, and ignoring
Try each. The one that's hardest to read is probably the one with the most in it for you.
1. Find your surrender zone before you delegate.
In certain styles of your work, your instincts are quick, and you catch mistakes fast there because you can feel when something's off. In another, the instinct isn't there at all. You'd rather not look. You delegate quickly because the work drains you. That second is your surrender zone.
Ask yourself: which kind of work do I send to AI first, with the least review? That's probably it. And that's exactly where AI's errors will hurt most, because you're not equipped to spot them.
For example, if you lead more from the analytical style, your surrender zone may be the relational work: how a decision will be received, what it'll cost in trust, what someone's actually saying behind what they said. AI gives you a plausible read, you accept it, and three weeks later you find out it missed what mattered.
If you lead more from the relational style, your surrender zone may be the numbers. AI gives you confident analysis, you run with it, and the math doesn't survive the board meeting.
Either way, your surrender zone needs more verification than your instinct will offer. If you're coaching a leader, the surrender zone question is often the first one to ask. It surfaces faster than most assessments.
2. Route the work, don't dump it.
Before you hand a task to AI, ask one question. Where in this task is AI strong, and where am I or my team needed?
Some categories are easy. Pattern detection, summarization, structured analysis. AI is genuinely good at these, and you can hand most of it over with light review. Drafting standard plans, generating SOPs, automating repeatable steps. AI assists well there too, with a second pair of eyes.
Other categories stay with you. Anything that touches trust, judgment, empathy, conflict, meaning. That work is yours. AI can generate words on those topics, but the words won't survive contact with the person they're for.
There's a third category that confuses most leaders. Strategy framing, possibility scanning, novel synthesis. This is partner work. You're thinking with AI rather than asking it.
Routing deliberately is the difference. Without it, whole categories of work get dumped into the chat window, and whatever comes back gets treated as the answer rather than the first draft of one.
3. Make AI argue with you. Not for you.
Pay attention to how you prompt. Most leaders use AI to confirm. To draft what they already thought, to polish what they already decided. AI is very good at converging. Hand it a problem, and it'll hand you back the average of the internet, said with conviction.
What it can do, and almost no one asks of it, is argue. Try this the next time you're about to make a real call: "Argue against my decision from the angle my team would underweight. What am I missing because I'm a [your dominant style] thinker?"
The first time, the answer will surprise you. The second time, you'll start asking the question yourself before AI does. That's the actual upgrade.
4. Use AI as a mirror. Not an oracle.
This is the practice almost no one runs. After you've made a decision, drafted a memo, or written up a meeting, hand the output back to AI with a different kind of question. Try: "Looking at what I just wrote, which thinking style was I leaning on? Which did I leave out? What would a leader from the opposite style have asked here that I didn't?"
What comes back is a sketch of your own thinking you'd never draw for yourself. The blind spots show up in soft outline. Your team can't give you this in the moment, because they're inside the decision with you. AI is outside it, and that's its rare value.
Try it on one real decision this month. Pay close attention to what surfaces. What you're getting back is your own thinking, seen from outside.
The through line is simpler than it looks. The leaders pulling ahead with AI share something simpler than prompt skill: they know how they think. They know where their thinking is strong, where it's thin, and what each looks like when AI gives it back amplified. They've moved from holding answers to hosting questions, and the teams around them feel the shift before the org chart shows it.
So pick one of these four practices. The one that made you uncomfortable to read is probably the one to start with. Try it on a real decision in the next few weeks and pay close attention to what surfaces. The measure that matters is whether you're sharper a year from now for having used AI at all.
If you're coaching a leader, ask which of these four they find hardest to even consider. That answer tells you most of what you need to know.
.png?width=6912&height=3456&name=Leadership%20Under%20Pressure%20Mastering%20Change%2c%20Complexity%2c%20and%20Uncertainty%20(1).png)
Ann Herrmann-Nehdi, Herrmann's Chief Thought Leader, hosts a free expert session on how to rewire leadership programs and make leadership ready for what's next. June 29 at 12:00 ET. Register here.
Share
