When I was asked which human skill matters more in the age of AI, I named two. Both are getting rarer. Both are getting more valuable. Neither is on the standard list.
On a panel last year, the moderator went around the table and asked everyone the same question. If you had to choose one human skill that becomes more valuable in the age of AI, what would it be?
I refused to choose one. I named two. Discernment and emotion. Both. Together. They don't work apart, and the operators who'll thrive in the next decade will have both, not one.
This isn't the standard answer. The standard answer is "creativity" or "empathy" or "critical thinking." Generic professional virtues that everyone agrees about and nobody knows how to hire for. Discernment and emotion are sharper than that. They're rarer. And right now, AI is doing two things at once that raises their value. It raises the floor on technical competence. It lowers the cost of producing things. The combination shifts where the work that matters actually sits.
Discernment is the ability to look at output that is technically correct and know that it's wrong anyway.
Not wrong in a fact-checkable way. Wrong in a way only judgment can name. Off-brand. Off-rhythm. Off-truth. The kind of wrong that another competent person would also flag, but couldn't articulate as a rule.
Discernment used to be an accelerant. You had a draft, discernment helped you make it better. Now it's the bottleneck. Because AI produces drafts at infinite speed and at uniform technical quality, the work has shifted from "produce the draft" to "decide which output is actually the right one." That's a discernment problem. There is no AI you can deploy to fix it. The model that wrote the output is exactly as confident in the wrong version as the right one.
The marketers I see compounding right now are the ones who can run twelve variants of a piece of content, look at all twelve, and say "this one. The other eleven are fine, but this one's the one." They can't always explain why. They don't need to. The discernment is the deliverable. AI gave them the twelve options. Discernment picks the one.
The marketers I see drowning are the ones who run twelve variants and ship whichever one Claude listed first, or whichever one was easiest to publish, or whichever one their boss reacted to most warmly in a Slack DM. They've outsourced the discernment to a defaulting process. The work is technically being shipped. The work is also bleeding distinctiveness without anyone being able to name the moment it happened.
You can't teach discernment in a workshop. You can only develop it by being made to make calls and being made to defend them. Most marketing teams stopped developing it because AI made the cost of "ship a draft" so low that the calls don't have to be made out loud anymore. That's a problem.
The second skill is harder to name and easier to dismiss. I'll try anyway.
Emotion is the ability to know how something is going to feel before it lands. Not what it'll perform. How it'll feel.
To the person reading the email. To the team holding the campaign. To the customer using the product. To the audience scrolling past the post. AI can model what people clicked on. It cannot model what something will feel like to be on the receiving end of, because feeling requires having a body and a history and the experience of being moved by something.
A marketer with emotion as a skill knows when a launch announcement is technically on-brief but emotionally cold. They know when an email subject line is structurally correct and reading-level right and still going to make the recipient feel like a number. They know when a campaign idea is exciting on a Notion page and dead on a billboard.
This is not "having empathy." Empathy is the standard answer and it's too soft. Emotion as a skill is closer to taste plus risk tolerance. It's the willingness to notice "this doesn't feel right" and to say so out loud, in a meeting, to people who have spreadsheets that say it should perform.
It's also the willingness to feel the work yourself. If a campaign doesn't move you, the chances of it moving the customer are low. Most teams trained themselves out of feeling their own work because feeling makes the meetings longer and the decisions slower. AI made it easier to skip the feeling step entirely. The output is fine. Ship it. Move on. The feedback loop that used to hurt is now smoothed out.
Emotion as a skill is the willingness to keep that loop unsmoothed. To care that something doesn't feel right even when nobody can prove it. To kill a campaign because it's cold even when it'd hit the targets. To rewrite the deck because it's mid even when the deadline is tomorrow. To insist on a launch image being made by a person because the AI version, which is technically perfect, doesn't move you when you look at it.
Discernment without emotion is technocratic. The leader who has it can pick the technically right option, but the picks lack warmth, and over time the brand starts to feel like a rules engine. Lots of holding-co strategy decks read this way. Right answers, dead.
Emotion without discernment is romantic. The leader who has it cares about the work but can't always direct that caring toward the option that actually wins. They love things. They defend things. The team admires them and ships uneven work.
The pair compounds. A marketer who has both can look at twelve AI-generated variants, feel the one that's alive, and articulate why it's the right call to a finance director who needs a defensible reason. They get the speed AND keep the work human. They don't need an AI checkpoint workflow for this part. They are the checkpoint.
This is who I'd hire for now if I were running a team. Not the person who's fastest with the prompts. Not the person with the cleanest portfolio. The person who can pick the right one from twelve, defend it without being defensive, and tell me when something I'm shipping is cold even though I can't see it yet.
If you manage marketers: stop measuring how many drafts the team produces. Start measuring how many they kill. The kill rate is your discernment proxy. A team that kills nothing is a team that's lost its discernment. A team that kills everything is paralysed. The right rate sits somewhere uncomfortable in the middle.
If you're a marketer: get into the habit of saying out loud "this doesn't feel right" without immediately defending it. The defense will come. The instinct comes first. Most of us trained that instinct out of ourselves to seem more rigorous. That training was wrong. The instinct was the rigor.
If you're hiring: stop interviewing people on prompts. Show them three variants of a piece of work and ask which is the one. Listen to the answer. Listen harder to how they got there.
Discernment and emotion. Both. Together. The two skills nobody's training for, and the two that compound.
Zara
Companion to "AI doesn't fail in marketing. Leadership does."
This essay sits inside the Marketing Prompt Hub: 30 tested Claude prompts across 8 marketing disciplines, 6 playbooks, 5 agent blueprints, full reporting kit. Pay once, own the files. No email gate.
See the hub →