I asked ChatGPT how to handle my team. Here’s what it missed.
Oct 19, 2026A while back, I asked ChatGPT how to handle a situation with someone in my team.
The advice I got was pretty helpful. It gave me simple steps and some really good questions to ask. But it never asked me a single question about me.
If I didn't know better, I would have walked into that meeting quietly confident and completely blind.
Last week I wrote about asking instead of telling. This week is about where we go for advice, and what that advice can't see.
What ChatGPT got right
I wanted to know how to have a difficult conversation with someone who wasn't open to feedback. They were very confident and quite self-assured, the type of personality that always tells but never asks.
ChatGPT gave me the simple steps of the process and the questions to broach the subject. That part it got right.
What it missed
It never asked about me. My attitude going in, or what I'd been projecting in my previous interactions with this person.
Had I gone in before in Parent mode or Adult mode? It didn't ask, so it didn't know. It didn't know what kind of rapport I'd already built with them, or how they might have felt after our previous conversations, based on how I was showing up.
So it only had half the story.
The real first step is reflecting on our own part. Then asking the team member some questions to understand their side, before we jump in to start fixing things.
The client who typed it in angry
One of my clients had taken feedback from a team member really personally. Sitting with that raw emotion and a bruised ego, he went straight to the keyboard and asked AI what to do.
In the heat of the moment, he painted a much gloomier picture of the team member than the reality. Based on what he put in, AI told him to start a heavy disciplinary process.
It validated his anger and his attitude, and it gave him more ammunition.
He took the steps with that angry energy. In the end it escalated, and the team member went to HR with a bullying claim against him.
It was resolved over time, through 1 to 1 coaching with him and some support for the team member. But it could have been avoided with a coaching session first, to prepare him for who he was being going into that situation.
No facts, just perceptions
When it comes to human behaviour, there are no facts. There are just perceptions.
AI can't read our mindset, our beliefs or the filters we're looking through. And we all have blinkers on about our own behaviour. We see other people's behaviour based on our own state in the moment, and we filter it through that.
Our mindset feeds our confirmation bias, and that shapes what we share when we type a problem into AI.
Tasha Eurich's research on self-awareness, in her book Insight, found that the people who rate their self-awareness highest are often the least likely to have high levels of it.
AI can only go off what we feed it. So the advice can be good. It's just working from our side of the story.
One thing to try this week
Before you act on any advice about someone in your team, whether it's from AI, a book or a colleague, start with you.
Ask yourself two questions. What have I contributed to this? What attitude am I bringing in?
Then ask them the questions first, before you tell them what you think.
See your own pattern
AI can give you good general advice. What it can't do is show you your own patterns.
That's what the Leadership Mirror is for. It looks at how you lead under pressure and shows you where your pattern is. It's free and takes about five minutes.
Prefer to watch or listen? Watch the video or listen to the podcast.
Have you ever asked AI about a leadership problem? What did it miss? Tell me in the comments.
Check out my On Line Course!
Take your Leadership skills to a whole new level, whether you are a seasoned professional or new to leadership.
Be Exceptional
The on Line Course that shares the secrets to how to show up as a Leader without the burnout and overwhelm.
