There’s an awkward question buried in all the talk about AI, and most leadership teams are quietly avoiding it. What happens when the youngest people in your business are more natural with the technology than the people running it?
Let me be clear about what I’m not saying. I’m not saying they’re more experienced, because they aren’t. I’m not saying they understand customers, risk, office politics or commercial pressure better than someone who’s spent 15 years learning all of it the hard way. That would be unrealistic. But something has changed, and it’s worth being honest about it.
I remember being that graduate. Eager, ambitious, slightly terrified, trying to prove I belonged and figure out how the real world actually worked, both at once. You came in knowing you had potential, and knowing all the experience sat around you rather than in you. So you watched people. You listened. You made your mistakes. You worked out how the senior people thought, and slowly you got some confidence. That was the deal. You came in to learn.
AI has changed that deal. Universities have unintentionally become some of the world’s largest AI experimentation environments. A graduate walking in today has probably already used AI to research, write, code, test, compare options and turn a blank page into something halfway decent. Hand them a P&L and they might blink at you. But sit them in front of a chatbot and a real problem and they’re off. They don’t understand your business yet. They might understand how to work with AI better than your business does. And that should make any leadership team stop for a second.
The job you start on has changed
Most careers were built on junior work. First drafts, research, admin, testing, notes, a bit of coding, pulling stuff together so someone more senior could take it the rest of the way. It wasn’t glamorous, but it did the job. It taught you how to think. It’s how you learned the difference between something that’s actually good and something that just looks good, which turns out to be most things.
That’s exactly the work AI is now doing. It’ll write the first draft, summarise the document, compare the options and generate a starting point in seconds. Which leaves a question a lot of businesses would rather not answer. If AI does the first job of a career, how does anyone learn to climb the ladder? This isn’t just changing graduate schemes. It’s changing how organisations develop judgement, experience and future leaders.
This isn’t a graduate problem, it’s a business problem. And there’s a bit of a contradiction in the market right now. Companies are shouting that they need more AI skills. The same companies are quietly working out how few junior people they can get away with. Those two things don’t sit together. Cut too many entry level jobs and you have to ask where the experienced people of the future are supposed to come from. Nobody gets to be senior without first being allowed to be junior, and you can’t build judgement if every safe chance to practise has been automated away.
So the real question was never whether AI would affect graduates. It already is. The question is whether anyone has thought properly about what replaces the old way of learning.
Being good with AI isn't the same as being good at the job
Here’s where I want to be careful, because it’d be easy to get carried away and claim graduates now know more than their managers. They don’t.
A 22-year-old can be brilliant with AI and still completely misread a client. They can hand you an answer that’s slick, confident and wrong in a way nobody spots until it lands on a customer’s desk three weeks later. They can move fast and miss the risk sitting right in front of them. They can give you something that looks finished and has never once been tested. Being confident with AI is not the same as having judgement, and anyone who mixes up the two is going to get caught out.
But being that comfortable with it is still worth a lot. Plenty of experienced people are cautious with AI because they’re trying to bolt it onto the way work has always been done. Younger people usually start somewhere else. They don’t see AI as a separate tool you go and pick up, they see it as part of how you start the work in the first place.
That’s a real difference, and the smart businesses won’t turn it into a scrap between age and experience. They’ll put the two together. Many organisations are beginning to embrace reverse mentoring, where junior employees share AI knowledge while experienced colleagues provide business context, governance and judgement. Senior people bring the context, the relationships and a real sense of where the risk sits. Younger people bring speed, curiosity and a lot less baggage. Get that mix right and it’s a serious advantage. Keep them in separate corners and you waste both.
The best ideas won't come from the top
One of the biggest mistakes with AI is assuming every answer has to come down from above. The board sets the ambition. IT picks the platform. The policy team writes the rules. Someone senior signs it off. Fine, all of that has its place. But the best AI ideas almost never start in that room.
They start with the person copying the same numbers into three systems every Monday morning and slowly losing the will to live. The one rewriting the same email for the twentieth time. The one clicking through five different places to answer one simple customer question. The person who knows a process is broken because they live in it every day. That person might well be a graduate. And if they’re already comfortable with AI, they’ve probably spotted a better way before anyone’s even asked.
The trouble is most businesses have no simple way to catch those ideas. So one of two things happens. People keep quiet, because they’re not sure what they’re allowed to do. Or they go off and sort it themselves with tools you don’t even know are in the building. That second one is where the risk starts, and notice the graduate isn’t the problem there. The lack of structure is. Give people no guidance, no safe tools and no way to raise a good idea, and you shouldn’t be shocked when it all goes underground. They’re going to use this stuff either way. The only real choice you’ve got is whether you help them use it well.
Maybe work needs to learn from them too
For years the question was whether graduates were ready for work. I felt that pressure myself. You wanted to show you belonged. You wanted someone to take a chance on you. You wanted to learn from the people who’d already done the hard miles. That still matters and it always will.
Maybe we’ve been asking the wrong question all along. Instead of asking whether graduates are ready for work, perhaps we should be asking whether work is ready for graduates.
Because this lot aren’t waiting for permission to think differently. They’re already using AI to learn faster, test ideas faster and get from question to answer faster than the business around them.
That does not mean leaving them alone with powerful tools and no supervision. If anything the opposite. They need mentoring more than ever. But it has to go both ways now. It can’t only be senior people explaining how things are done here. It also has to leave room for a graduate to say, gently, that “how things are done here” stopped making sense a while ago.
That’ll make some leaders uncomfortable. Good. AI should make us a bit uncomfortable, not because it kills off the value of experience, but because it makes us ask an honest question about where the value actually comes from now. Is it knowing the answer? Or is it knowing how to ask the right question, check whether the answer holds up, spot the risk and apply some judgement? I think it’s tipping towards the second.
What a graduate scheme should actually look like
The lazy answer is to throw a few AI tools at your graduates and hope for the best. That won’t work. You’ll just get faster chaos.
The better answer is to build early career development around the world these people are actually walking into.
The organisations getting this right will:
- Give graduates approved AI tools rather than leaving them to find their own.
- Pair them with experienced mentors who challenge their thinking, not just their outputs.
- Encourage experimentation within clear governance, so people can innovate safely.
- Reward ideas that solve genuine business problems, not just demonstrate what AI can do.
- Teach judgement alongside AI capability, because knowing when not to use AI is just as valuable as knowing how to use it.
None of that removes the need for experience. It makes experience more valuable. Graduates still need to learn how organisations work, how customers think, how risk is managed and how good decisions get made. But they also bring perspectives and capabilities many organisations are only just beginning to develop.
Stop treating graduates like they’ve got nothing to offer until you’ve trained them. They’re still learning, obviously. But some of them are already highly capable with AI in ways the business genuinely needs. The organisations that recognise that early won’t just build better graduate schemes – they’ll accelerate AI adoption across the business.
So the smartest thing you could do this quarter probably isn’t buy another tool. It’s walk over to the newest person on your team and ask how they’d do your job. You might not like the answer. That’s rather the point.
This is the second article in a series from MTI’s Chief AI Officer exploring what successful AI adoption really looks like, from governance and security to agents, data readiness and measuring business value.
Next in the series: AI Is Not A Technology Revolution. It’s A Human One.
Continue the conversation
The questions raised in this article are ones we’re discussing with organisations every day. From AI readiness and governance to practical adoption and security, MTI helps organisations move beyond the hype and build AI that delivers real business value.
About The Author
