A Swedish fintech company made a bold announcement in early 2024. Its AI assistant would take over all customer conversations. It had handled about 2.3 million conversations, 75% of the company’s customer chats. The company said the bot did the work of 700 agents. So it was getting rid of human agents for good. The CEO was very proud. He had frozen hiring. He said, publicly, that AI could do all the jobs of people. Then reality set in. By 2025, he told Bloomberg that while they saved costs, the AI push was hurting service quality for their human customers. So they are back to human agents now. “From a brand perspective, a company perspective, I just think it’s so critical that you are clear to your customer that there will always be a human if you want,” the CEO said. I love this true story. The AI advantages were good. And so was the damage.

A lot has changed since 2024.

Many AI agents are more capable now. Most companies are going all in. And it’s not just customer conversations. The bots are coming for all types of jobs. The career anxiety people feel makes sense. The future of work is terrifying. Start with what’s happening to young workers. Stanford economists have tracked this in a study they call

The August 2026 update says employment in AI-affected fields is down 19% for young workers compared to more AI-resistant careers. A year earlier, that gap was only 13%. The researchers said the gap was driven mostly by weaker hiring. People were not being laid off. They were not quitting either. Nobody gets fired.

Except the door never opens for anyone.

Now the part that matters most for you. The same study found something interesting. Jobs that rely more on experience and know-how grew faster among mid-career and senior workers. Not so much for entry-level careers. Jobs that require just pure know-how were being eaten by AI. PwC did a similar study; they reported that AI-exposed roles are seven times more likely to list “traditionally senior” skills. Employers now want juniors who can act like seniors. In another study, PwC’s AI Jobs Barometer came across a useful insight. Workers who are so good at making sense of AI outputs, how it makes company money or saves costs, earn 56% more than their peers. The bar is being raised. Knowing how to use AI is now basic. Also, the World Economic Forum ranks.

among today’s core skills, with 50% of employers identifying them as essential.

Enough data.

Now, what?

Practical skill stacking is what can save you right now.

Start with judgment. Knowing what is good, better, or best for your employer or your own work. The philosopher Michael Polanyi said, “We can know more than we can tell.” Of course, good judgment comes from thousands of hours of seeing good and bad work side by side. AI struggles with practical real-world sense and intuition. These skills are hard to program into a machine. That’s where you come in. Super-intelligent machines can give you dozens of options. That’s a start. Now, who gets to pick what works or what should be applied? The one who has seen enough to know from experience. The indispensable worker who can’t be replaced. That person is the bottleneck who gets paid now and in the future. HBS Professor Karim Lakhani says, “The lesson? AI is not a replacement for judgment,” “Knowing where to apply it, and where not to, is now a critical leadership skill.” Employers now want people who make practical sense of all that knowledge AI is given their company.

Judgment is experience compressed into instinct.

And it’s rare.

The new intelligence tools can give you all the knowledge you want. It probably can’t tell you what could damage your existing work out there, make you customer angry or violate unspoken rules of your industry. It’s being trained to get smarter. But you still have to make the final call. Building judgment takes work. You have to act by doing. Make decisions and live with the consequences. You have to be able to take the next best action even when the stakes are high. And learn from it.

There’s no shortcut.

Another vital rare skill is knowing how to think through beautiful and thought-provoking questions. And using them to get what you want. Picasso is said to have told an interviewer in 1964 that computers are useless because they can only give you answers. I love that. Knowledge is cheap now. They cost close to nothing. So what questions can get you the answer you need? I need a marketing plan to attract buyers won’t get you anywhere. “Why did our best customers stop buying in March? And what would have to be true for them to come back?” Take it to a whole practical level. That’s what you want. Specifics. But to reflect on the right questions, you need to understand the business, the customers, and the pain. An AI can rarely help with that. It can give you knowledge it’s being trained on. But the instinct to know what works in the real world must come from you.

I like what Einstein said, “If I had an hour to solve a problem and my life depended on the solution, I would spend the first 55 minutes determining the proper question to ask, for once I know the proper question, I could solve the problem in less than five minutes.”

The next thing smart people are honing these days is originality and taste. Right now, AI is flooding the market with knowledge. Mediocre creativity won’t cut it if you want to stay ahead of something that’s changing too fast to make sense of. Taste is the difference between a product that informs and one that makes people want to buy. AI can generate a thousand variations. It cannot tell you which one will make your customer feel something. There’s a practical reason taste matters more now. When everyone has access to the same tools, the differentiator is no longer production capacity. It’s editorial capacity. The ability to say subtract. Reject what doesn’t make people connect or engage. And to look at dozens of good options and choose the one that fits a vision no one else can see.

Taste is how you don’t fall for the knowledge trap.

In the words of physicist Richard Feynman, “The first principle is that you must not fool yourself, and you are the easiest person to fool.” Taste is looking at ten high-quality outputs and immediately spotting the one that makes the most human sense. If you cannot tell bad work from great work without a team, committee, or a metric, you are vulnerable. If you have undeniable taste, AI just becomes your hyper-efficient intern.

AI makes it harder to hone taste.

It can easily convince you of its argument. It gives you what you want with confidence, even when it’s wrong. People mistake it for the expert. It doesn’t make you want to think. Your brain loves that. It wants to say yes and move on. The people who check, think through what they get, and decide with caution are rare. Taste as a skill is boring. That’s why it’s rare. It takes work. But you can build it with a simple rule. For anything that matters, verify the sources, the numbers, and how practical it is. Make it a habit before you make it a system.

Be the human in the room.

Go back to the Swedish fintech company.

Why did they bring people back? Well, customers want to talk to people who understand the details. Who take human responsibility. Someone who could hear frustration, think in real time, and do the human thing. Or say “I’m sorry, let me fix this for you.” That alone creates better connection and trust. And it’s not just a soft skill. It’s a human thing to do. Trust is built face to face, in follow-up calls nobody asked for. In taking care of people outside the code. Be more curious about what works for the people who matter. The WEF’s

ranks analytical thinking as the top core skill. Seven in ten companies call it essential. The report also lists creative thinking, resilience, flexibility, agility, and curiosity as rising in importance. These are human skills.

You can keep honing your people skills. Pick three people this month and do something for them that takes human effort. Make an introduction. Send useful feedback. Be there for them. Be indispensable in your team. Be the person your boss can’t imagine losing. Be the person others rely on.

Also, be comfortable with taking initiative.

Be more curious about what could work or save time. Curiosity is refusing to accept the first answer. And thinking to yourself, “what would have to be true for this to be wrong?” Curious people learn. They test and do things. And don’t wait for permission to figure things out. Most people spend their entire careers waiting for an assignment. They want a brief, a spec sheet, or a green light. The rare skill today is seeing a problem, diagnosing the cause, and coming up with what works before you get the green light. When algorithms handle routine execution, the most valuable person in the room is the one who says, “follow me, I figured out a vital answer for this.” Irreplaceability is your personal responsibility. One of the biggest risks to your career is “curiosity offloading.” Letting algorithms decide what deserves to be looked into. If you are getting compelling answers without creative effort, you lose the productive struggle that builds judgment. The rare skill is disciplined wonder.

Use AI to go further and wider.

But not to avoid thinking for yourself. Most people use AI to get fast answers to shallow questions. Machines solve for the constraints you give them. Your job is to decide whether the constraints are practical in the first place.

Before you go, remember, these skills are not taught in most schools. You can’t certify taste, judgement, or holy curiosity. They’re built through experience, reflection, and a willingness to be wrong.

The good news is that AI used at the right time, for the right task, can give you more time to build them. You can get help for the routine execution. What’s left is the part that matters. Deciding what to make, judging whether it’s right, reflecting on better questions. And owning the outcome. Nobel laureate Daron Acemoglu once said, “Don’t be taken by the hype.” Think how your human skills can be better deployed. There’s more to do for skilled workers. AI is not getting rid of careers anytime soon. But it’s destroying the careers that were built on execution alone. The rare skills, judgment, taste, curiosity and connection are not new. They’ve always mattered. But they’re even more important right now. None of them is a tool. You can’t buy them. You have to work for them.

You grow them, slowly, by doing things that feel inefficient.

That’s exactly why they’re rare, and why they’ll stay rare.

Kierkegaard wrote that life can only be understood backwards, but it must be lived forwards. Careers are like that. You’ll have to work through it forward, with imperfect knowledge, like everyone before you. Seneca said, “It is not because things are difficult that we do not dare; it is because we do not dare that they are difficult.” So here’s my question. Which of these skills will you practise this week? Pick it. Do it badly. The process matters more. Repeat to hone. The industrial economy rewarded consistency. The information economy rewarded knowledge acquisition. Now, the AI economy, I suspect, will reward judgment, taste, curiosity, initiative, and making the most of your ability to connect or build trust with other humans.

These are your economic value.

The things you alone can bring to the table. You can’t outthink a super intelligent machine. You just need to stop acting like one. And start taking responsibility for the results only a human can deliver.