The New Smart

In short: As AI makes technical and analytical work easier to produce, emotional intelligence, the ability to read people, earn trust and move a group toward a decision, becomes a scarcer and more valuable skill, alongside judgment.

For most of my career, there was a clear picture of who the smartest person in the room was. It was the one who could do the hard technical thing nobody else could. Write the code. Build the model. Solve the problem on the whiteboard while everyone else was still reading it. That person got the title, the raise and the benefit of the doubt.

I'm watching that picture change in real time.

The hard technical thing is getting easy. Not trivial, and not for every problem, but easy enough that it no longer marks someone out. A capable model can write working code from a plain-language description. It can draft the analysis, structure the argument and produce in an afternoon what used to take a specialist a week. The skills we spent decades treating as the definition of intelligence are becoming something you can ask for.

When a skill becomes something you can ask for, it stops being what makes you valuable. Something else takes its place. I think that something is emotional intelligence: the ability to understand people, read a room, earn trust and move a group toward a decision. Long treated as the soft skill, the nice-to-have, it's becoming the hard one.

How Did IQ Become the Measure of Smart?

Our idea of what counts as "smart" has always followed the economy.

In the industrial era, intelligence meant something close to reliability and discipline. In the information era, it came to mean analytical horsepower: the ability to manipulate symbols, follow logic, hold complexity in your head and produce correct answers. Standardized tests measured it. Elite schools selected for it. The highest-paid jobs rewarded it. Coding became the clearest modern symbol of it, a skill that looked like pure reasoning and paid like it, too.

That definition of intelligence was always incomplete. But it was a very good description of what was scarce. Machines couldn't reason in language. Computers did exactly what they were told and nothing else. So the people who could translate human intention into precise machine instruction held enormous leverage.

That's the job language models are now taking on. They translate human intention into code, into analysis, into structured output. The bottleneck that made the classic technical mind so valuable is opening up.

What Skills Is AI Making Abundant?

"Coding is irrelevant" is the kind of claim that ages badly. Someone still has to understand systems deeply, know when the generated code is wrong, and design architectures that hold up. The best engineers I know are becoming more valuable, not less, because they can now direct ten times the output.

But look at what's actually happening to the market for the skill itself. In one controlled study, developers using GitHub Copilot finished a coding task about 56% faster. At a large customer support operation, AI raised productivity about 14% overall and about 34% for the newest workers. In a study of BCG consultants, AI users worked about 25% faster with about 40% higher quality, and the biggest gains went to those who started below average. AI is closing the gap between the average and the expert. That's the signature of a skill becoming a commodity. When the floor rises that fast, being good at the thing stops being the differentiator. Being good at everything around the thing starts to matter instead.

The same pattern runs through every task we used to call "high IQ." Summarizing research. Building a financial model. Writing a clean argument. Solving a well-defined problem. These aren't going away. They're becoming inputs. Anyone with access to a good model can produce a competent version.

And competent, it turns out, was most of what we were paying for.

What Skills Stay Scarce?

Some things a model can't do, at least not in any way that counts.

It can't sit across from a client who is scared about their business and make them feel understood enough to tell you what's really going on. It can't notice that the quietest person in the meeting disagrees, and make it safe for them to say so. It can't hold a team together through a bad quarter. It can't tell when a stakeholder's objection is really about their status, not the plan. It can't earn the kind of trust that makes people follow a recommendation they don't fully understand yet.

People read people from inside a body, a history and relationships that can be lost. That's where emotional intelligence lives. It's prediction about people, built from years of stakes: being hurt, being forgiven, watching what happens when you say the wrong thing at the wrong moment. A model can describe empathy beautifully. It has never needed anyone.

Emotional intelligence is usually described as a cluster of abilities: noticing emotion in yourself and others, understanding what's driving it, and managing it well enough to act wisely. Strip away the jargon and it's simpler than that. It's the ability to work with people as they actually are rather than as you wish they'd be.

That skill was always valuable. What's changing is that, along with judgment, it's becoming what only people can bring. Judgment decides what's right. Emotional intelligence gets people to act on it.

What Does the Evidence Show?

This shift didn't start with AI. It just accelerated.

Economists studying the labor market found that, over the past few decades, jobs requiring strong social skills grew faster and paid better than jobs requiring math skills alone. The jobs that held up best were the ones that needed both. Meanwhile, roles that were highly technical but low on human interaction were among the most exposed to automation even before generative AI arrived.

Employer surveys tell a similar story. When companies are asked which skills they expect to grow in importance, analytical thinking still ranks high, but so do resilience, flexibility, curiosity, leadership and social influence. Those are not the skills we built our schools around.

And I see it in my own work every week. In strategy work, the deck is rarely the hard part. The hard part is the moment in the room when a stakeholder goes quiet, and you have to decide whether that silence means agreement or a problem nobody wants to name. The projects that fail rarely fail because nobody could build the thing. They fail because people weren't aligned, someone didn't feel heard, a decision got made in a room where the real objection never surfaced. The technical work is almost never the hardest part. The human work almost always is.

What Does This Mean for Organizations?

If this is right, a lot of how we hire, promote and develop people is pointed at the wrong target.

Hire for judgment and relationships, not just output. If a model can produce the output, the person's value is in what they decide, who they can influence and whom they can bring along. Interview for that directly.

Promote the people who make rooms work. Every organization has people who quietly make everything around them better: they resolve conflict, translate between teams, make others feel capable. We've historically undervalued them because their contribution is hard to measure. It's about to be the most important contribution there is.

Train emotional skills like you train technical ones. We treat empathy as a personality trait, something you either have or don't. It's more learnable than that. Listening, giving hard feedback, reading a stakeholder, staying steady under pressure: these can be practiced and coached.

Redesign technical roles around the human edge. The engineer of the next decade spends less time writing code and more time understanding what people need, explaining trade-offs and making judgment calls. Build roles, and career paths, that reward that.

Watch for the new blind spot. Teams that lean heavily on AI can get very fast and very lonely. Fewer conversations, less friction, less of the messy human contact where trust actually gets built. Protect that contact deliberately.

What Are the Objections?

There are good reasons to push back on all of this.

First, emotional intelligence is harder to define and measure than IQ, and some researchers argue its popular version overstates what it predicts. That's fair. The case here doesn't rest on any particular test. It rests on the observation that working well with people is becoming the scarcer skill.

Second, models are getting better at the interpersonal side, too. They can be patient, attentive and remarkably good at sounding understanding. Some people already find them easier to talk to than other people. I don't dismiss that. But sounding understanding and being accountable for a relationship are different things. Trust, at least in work that matters, still runs between people who can be held responsible.

Third, technical depth isn't disappearing. Someone has to know whether the generated system is safe, correct and well designed. The likelier future isn't EQ instead of IQ. It's that raw analytical ability stops being enough on its own, and the people who combine judgment with human skill pull far ahead.

The Room Is Changing

For a long time, we told a certain kind of person that their gift was secondary. The one who could tell how everyone was feeling. The one who knew what to say in a hard conversation. The one who made the team work but couldn't point to a single thing they'd built. We called it soft.

It was never soft. It was just hard to measure and easy to take for granted.

The smartest person in the room is no longer the one with the answer. The answer is a prompt away. The smartest person in the room is the one who knows what the people in it actually need, and how to get them there together.

That's the new smart.

Frequently Asked Questions

What is emotional intelligence? The ability to notice emotion in yourself and others, understand what's driving it and manage it well enough to act wisely. Put simply, working with people as they actually are.

Will AI make technical skills irrelevant? No. Deep technical judgment still matters. What's changing is that raw technical output is no longer enough on its own to stand out.

Can emotional intelligence be learned? Largely, yes. Listening, giving hard feedback, reading stakeholders and staying steady under pressure can all be practiced and coached.

Evan Prout is a strategist in Grand Rapids, Michigan, who writes about how organizations think and decide with AI.