Emotional Intelligence: The Developer Skill AI Can't Replace

Hi lovely readers,

A few months ago I wrote about how AI is making English a must-have skill for developers. The idea was simple: when you spend more of your day describing problems instead of typing code, the way you communicate becomes much more important.

Recently I sat in on a talk by Manuela Beltrami at a women in tech event for Rotterdam University of Applied Sciences that I help organize. Her talk was all about failure, and one slide stuck with me. It called emotional intelligence your superpower in the age of AI. That slide kept going round in my head, and it made me realize there was a second half to my English post that I had not written yet.

Writing a clear sentence is one thing. Knowing what the other person actually needs to hear, how they are feeling, and how you are feeling yourself is something else. That something else is emotional intelligence, and I think it might be one of the most underrated developer skills right now.

What is emotional intelligence?

Let me try to explain this term in the easiest way possible. Emotional intelligence is the ability to notice and understand emotions, both your own and other people's, and to use that understanding to handle situations well.

The term became popular through psychologist Daniel Goleman and his 1995 book Emotional Intelligence. He breaks it down into five components:

1. Self-awareness: recognizing your own emotions and how they affect your work.
2. Self-regulation: managing those emotions instead of reacting on impulse.
3. Motivation: being driven by more than external rewards.
4. Empathy: understanding what others feel and need.
5. Social skills: building relationships, communicating, and handling conflict.

The first three are about you. The last two are about the people around you. You will see all five come back in this post.

At work, it looks like this:

• Noticing that you are frustrated before you send that message in Teams or Slack
• Understanding that a stakeholder who keeps changing their mind might be under pressure from their own manager
• Seeing that a colleague has gone quiet in meetings and asking if they are okay
• Giving feedback in a way that helps someone grow instead of making them defensive

None of this involves code. But all of it affects how well software gets built.

AI does the typing, not the understanding

AI tools are getting better every day. They can generate an endpoint, write tests, and explain an error message. Tasks that used to take an hour now take a few minutes.

But if you look at why projects actually go wrong, it is rarely because someone could not write a for loop. It is because the team built the wrong thing. Because two people misunderstood each other. Because someone was afraid to say they were stuck. Because a disagreement about architecture slowly turned personal.

AI does not fix any of that. It can transcribe a meeting and maybe even guess at the mood in the room, but it misses the context that gives those signals meaning. It does not know that the client who says "that's fine" said exactly the same thing right before the last project went sideways. It does not know your colleague just had a really hard week, or that your stakeholder's manager is breathing down their neck. It can sound friendly and understanding, but sounding kind is not the same as actually understanding.

So as the technical part of our work gets faster, the human part becomes a bigger piece of what makes you valuable.

Empathy helps you find the real problem

The more clearly you describe a problem, the better the answer you get from AI. But there is a step before that. You first need to understand what the problem actually is.

That is where empathy comes in. When a user says "this page is annoying," the real problem might be that the button is hard to find, that the page is slow on their phone, or that they are simply stressed and this page is the last thing standing between them and going home. You only find out by putting yourself in their shoes and asking the right questions.

Accessibility is empathy you can ship

This is also why I care so much about accessibility. For me, it is the most practical form of emotional intelligence in our field.

Accessibility means thinking about people who are not in the room. The person using a screen reader. The person with shaky hands who struggles with tiny buttons. The person reading your UI in their second language, or on a bad day when their focus is gone.

AI can help here. It can suggest alt text, point out missing labels, or check your color contrast. But it only does that when someone asks. If you never thought about those users in the first place, no tool will do that thinking for you. The empathy has to come first, and the tooling comes after.

Managing your own emotions matters too

Emotional intelligence is not only about other people. It is also about you.

Let's be honest, working in tech right now can be stressful. There is a lot of noise about AI replacing developers, the job market is harder than it used to be, and it is easy to feel like you are always behind on the newest tool. On top of that, working with AI can be frustrating in a very specific way. You explain something three times, it confidently gives you the wrong answer again, and you feel your patience disappear. By the end of the day, these situations can really get to me.

Being aware of those feelings helps a lot. When you notice "I am frustrated right now," you can take a break instead of accepting code you do not understand just to be done with it. When you notice "I am anxious about my future," you can talk about it with someone instead of letting it quietly stress out your work and your sleep. In Goleman's terms, noticing the feeling is self-awareness, and choosing what to do with it is self-regulation.

It also helps to know what keeps you going. If your only motivation is keeping up with every new release, you will always feel behind, because there will always be another one. Maybe for you it is curiosity, or the moment a hard concept finally clicks, or building something that actually makes someone's day easier. That is what Goleman means by motivation: being driven by something deeper than the next shiny tool or a pat on the back. It is also what gets you through the weeks when the hype is exhausting.

Failure builds emotional intelligence

Here is something Manuela's talk made me realize: every time something goes wrong, you are training your emotional intelligence, whether you notice it or not.

Dealing with disappointment when a release breaks. Asking for help when you are stuck, even though it feels embarrassing. Supporting a colleague who just took down production. Reading the room in the meeting after something went wrong. All of that builds the same muscle.

When I look back at my first years as a developer, the moments that taught me the most about myself were not the successes. They were leaving a job that undervalued me, working somewhere that just did not click, and getting let go before I had even started. None of that felt good at the time. But it taught me how to handle disappointment, how to stop blaming myself for everything, and how to recognize what I need.

AI can help you avoid some mistakes. It cannot go through them for you. And the growth that comes from getting through a hard moment, and helping others get through theirs, is still something only you can do.

Emotional intelligence in PR reviews

AI means more code gets written, which also means more code needs to be reviewed. And reviewing is one of the most emotional parts of our job, even if you think that it is not. This is where Goleman's social skills come in.

When you review a pull request, try to be curious instead of harsh. "Can you walk me through why you chose this approach?" lands very differently than "Did AI write this?" Even when AI did write it, the person who opened the pull request is still a colleague, and they are still the one who will learn from your feedback.

The same goes for receiving feedback. Your code is not you. A comment on your pull request is about the code, not about your worth as a developer. That is easy to say and hard to feel, and learning to feel it is part of emotional intelligence.

Emotional intelligence is not the same as being outgoing

I want to be clear about this one, because I think a lot of people get it wrong. Emotional intelligence does not mean being the loudest person in the room, loving networking events, or being friends with everyone.

As I shared in an earlier post, I struggle with busy social events and open offices. That does not make me bad at understanding people. Quiet people are often very good at this, because they spend more time listening and observing. And knowing your own limits, like knowing when you need a quiet space to recharge, is emotional intelligence as well.

Emotional intelligence and neurodiversity

There is a stereotype I want to push back on: that neurodivergent people, for example autistic people or people with ADHD, are "bad" at emotional intelligence. That is not true, and it is not a helpful way to think about it.

A lot of what we call emotional intelligence is based on one specific way of communicating. Making eye contact, reading between the lines, picking up on tone of voice, understanding what someone means when they actually say something else. If your brain works differently, those signals might not come naturally. But that does not mean you do not care or do not feel. Many neurodivergent people feel emotions very strongly, including other people's.

There is a term for this: the double empathy problem. The idea, which comes from autism researcher Damian Milton, is that when an autistic and a non-autistic person misunderstand each other, the problem goes both ways. It is not that one person lacks empathy. It is that two different communication styles meet, and both sides find the other hard to read.

That changes what emotional intelligence looks like in a team. It is not only about reading others well. It is also about making yourself easy to understand, and accepting that people show things in different ways.

In practice, that can look like this:

• Say what you mean. "Can you fix this before Friday?" is much clearer than "It would be nice if someone looked at this soon."
• Do not read tone into short messages. A one-word "ok" in Slack is usually just an "ok."
• Ask how people like to work instead of assuming. Some people prefer a quick call. Others want things in writing so they can process them in their own time.
• Make room for different energy levels. Cameras off, quiet spaces, and meetings with a clear agenda help more people than you might think.

Interestingly, AI can help here too. Some neurodivergent developers use AI tools to check how a message might come across, or to ask "what could this person have meant?" That is a nice example of AI supporting human connection instead of replacing it. Still, it is a tool, not the solution. The real work is people meeting each other halfway.

Emotional intelligence was never about everyone communicating the same way. It is about understanding that we do not, and making it work anyway.

Teaching and mentoring are changing

As someone who teaches and mentors, I notice that what learners need from me is changing. They can ask an AI to explain a concept at 2 AM, as many times as they want. What they cannot get from it is someone who notices they are overwhelmed, who tells them that being confused is a normal part of learning, or who helps them believe they can actually do this.

Knowledge is easier to get than ever. Encouragement, trust, and feeling seen are not. Those still come from people.

How to get better at it

The good news is that emotional intelligence is not something you either have or do not have. You can practice it, just like any other skill.

• Listen to understand, not to reply. In your next meeting, try to repeat back what someone said in your own words before you answer.
• Ask one more question. "What would make this a success for you?" often tells you more than the whole requirements document.
• Pause before you send. Reread your message as if you are the person receiving it. How does it sound on a bad day?
• Name what you feel. Just saying "I am frustrated" or "I am nervous" to yourself makes it easier to handle.I sometimes even say it out loud at conference talks in front of an audience.
• Ask, do not assume. How someone communicates tells you about their style, not about how much they care.
• Ask how you come across. A trusted colleague can tell you things about your communication that you will never notice yourself.

That's a wrap!

AI is changing how we write code, and that is exciting. But software is still built by people, for people. The better you understand the humans around you, and yourself, the more valuable you become, no matter how good the tools get.

Or, as Manuela put it on that slide: "AI is a tool. You are the strategy."

Thanks for reading! If you have thoughts, disagree with me, or want to share your own experience, feel free to leave a comment or reach out on my socials. I am always happy to talk about this stuff.