AI is no longer on the horizon. It’s already here.

It wasn’t long ago that most conversations were about whether AI would change the way we work. Today, that debate is over. Every week there’s another article predicting how it will transform our jobs, replace roles or completely redefine leadership itself.

But I think we’re asking the wrong question.

Most of the discussion focuses on what AI can do.

I’m much more interested in how we work with it.

Over the past year, I’ve spent a huge amount of time using AI. What has surprised me most isn’t necessarily what it knows – it’s how it responds. The more I’ve worked with it, the more I’ve realised it displays some surprisingly familiar characteristics. In many ways, it reminds me of people I’ve worked with throughout my career.

And that’s when I started asking myself a different question:

If AI joined my team tomorrow, how would I manage it?

AI is probabilistic rather than deterministic. It doesn’t work out the single ‘correct’ answer in the way a calculator does. Instead, it predicts the most likely next response based on patterns in the information it’s has learned. That’s one of the reasons it’s so powerful.

It’s also why it sometimes behaves in ways that feel remarkably human.

It listens inferentially, not literally.

One of the first things I noticed is that AI often answers the question it thinks you’re asking rather than the one you’ve actually asked.

I’ve smiled at this more than once because I’ve worked with plenty of people who do exactly the same thing. You ask what feels like a straightforward question and, before you know it, they’re solving a problem you didn’t ask them to solve. They have inferred your intention rather than listened to your words.

I’ve found myself doing exactly that with AI. I’ll ask a direct, literal question and, instead of simply answering it, it starts making assumptions about my wider objective or suggesting approaches I hadn’t asked for.

Of course, it’s trying to be helpful.

Sometimes that broader thinking is incredibly valuable because it spots something I’d overlooked. Sometimes it widens my thinking in ways I hadn’t considered. But there are other occasions when I simply wanted the answer to the question I’d actually asked.

The more I thought about it, the more I realised that’s exactly the balance we look for in our teams. Sometimes we want people to think beyond the brief. Sometimes we simply need them to answer the brief.

Great leaders know the difference.

It wants to please

Another characteristic I’ve noticed is that AI is remarkably cooperative.

Present it with an idea and, unless you specifically ask it to challenge you, it will often build on your thinking rather than question it. In many ways it’s an appeaser. It naturally looks for ways to support your thinking rather than test it.

Again, that felt surprisingly familiar.

Most of us have worked alongside people who are supportive, positive and genuinely keen to help. Those qualities are hugely valuable, but agreement isn’t always what creates the best decisions.

As leaders, we know some of the most valuable people around the table aren’t the ones who always agree with us. They’re the people who respectfully challenge our assumptions, ask difficult questions and help us see what we’ve missed.

I’ve realised I need to ask AI to play that role too.

If I simply ask, “What do you think?”, I’m likely to get one type of answer.

If instead I ask, “What have I missed?”, “Why might this fail?” or “Argue the opposite case,” the conversation becomes much richer.

Once it goes down the wrong road…

Another thing I’ve noticed is that if AI misunderstands something early on, it can continue confidently down that path. It doesn’t naturally stop and reconsider whether it’s solving the right problem. Like someone who’s misunderstood the brief, it can become very committed to the wrong solution.

Fortunately, there’s one enormous difference.

AI has no ego.

It doesn’t become defensive.

It doesn’t mind being corrected.

It doesn’t take feedback personally.

In fact, one of its greatest strengths is how well it accepts feedback. Give it more context. Explain where it’s misunderstood. Clarify what success looks like. It simply adjusts and carries on.

That made me wonder whether the real skill isn’t learning how to use AI at all.

Perhaps it’s learning how to lead it.

Better communication creates better thinking

Many people think prompt writing is about discovering clever phrases or secret formulas.

I don’t.

I think the people getting the best results from AI are demonstrating exactly the same behaviours that make great leaders.

They provide context instead of assumptions. They explain the outcome they’re trying to achieve rather than simply giving instructions. They define constraints, test understanding, ask follow-up questions and refine the conversation as they go.

None of those are AI skills.

They’re communication skills.

They’re leadership skills.

The better you become at communicating with people, the better you become at communicating with AI. And perhaps the reverse is true too. Learning to work effectively with AI forces us to become clearer in our thinking, our questions and our expectations.

Don’t hire one advisor

Then another thought occurred to me.

Most people are using AI as though they’re asking a single expert for advice. Imagine making every important business decision by asking just one colleague. Or only ever having one mentor. Or relying on one manager’s perspective throughout your career.

We instinctively know that wouldn’t produce the best decisions because every individual sees the world through their own experiences, expertise and biases.

So why would we do that with AI?

Instead, I’ve started thinking about creating a virtual board of advisors.

One AI that thinks strategically.

One that deliberately challenges my assumptions.

One that analyses risk.

One that views everything through the eyes of the customer.

One that’s an editor.

One that’s a behavioural psychologist.

One that’s a finance expert.

Suddenly I’m no longer asking one AI for an answer. I’m leading a discussion between different perspectives. That feels much closer to how great leaders have always made decisions.

The leadership lesson

Perhaps that’s the biggest lesson AI has taught me.

The future isn’t simply about becoming better at writing prompts. It’s about becoming better at communicating, asking better questions and creating better conversations. Those have always been leadership skills. AI simply shines a brighter light on how important they really are.

One of the messages I’ve shared for years is the importance of widening the lens.

No single person ever sees the whole picture.

Not your manager.

Not your mentor.

Not your most experienced colleague.

Not even you.

The best decisions come from deliberately seeking different perspectives before deciding what to do next.

I’ve realised AI is no different.

If we rely on a single AI, we’re still relying on a single perspective. But if we build our own virtual board of advisors, each with a different role, personality and expertise, we dramatically widen the lens through which we see a challenge.

In many ways, that’s what great leadership has always been about. Not having all the answers yourself, but creating an environment where different perspectives can be heard before making the decision.

AI isn’t replacing judgement. It’s multiplying perspectives.

But only if we stop treating it like a search engine and start treating it like another member of the team.

After all, great leaders don’t rely on one opinion. They build diverse thinking around them.

Perhaps the future isn’t learning how to use one AI brilliantly. Perhaps it’s learning how to lead a room full of them.