For most of my career in engineering and construction, there was an unspoken ranking of what mattered. Technical ability sat at the top. It could be measured, tested, and billed. Everything human about the work sat below it, filed under a label that did lasting damage: soft skills.
I believed the ranking myself for a long time. I came up as a civil engineer, and the message was consistent: Master the technical work first. The people part can wait.
Then I moved into executive leadership, and the ranking collapsed. Almost nothing that determined whether my organization succeeded was technical.
Success turned on whether people trusted each other enough to tell the truth, whether hard conversations happened early or festered for months, and whether leaders could make sound calls once the data ran out.
The skills I had been told were secondary turned out to be the whole job. AI is now forcing that same realization on every organization at once, and far faster than my career forced it on me.
Consider what these tools do best: producing a first draft, analyzing a dataset, building a model, and summarizing a hundred pages. This is precisely the work we spent decades treating as the most valuable, and it is the work AI is learning fastest.
Meanwhile, the abilities no tool can replicate are the ones we dismissed. Rebuilding trust after it has been damaged. Reading what a silent room is telling you. Choosing well when every option carries risk and information has run out. Helping people act when no answer is clean.
If you work in communications or HR, or lead employee experience, you know this dismissal from the inside. Your domain of trust and human connection spent decades classified as a support function. Helpful, but optional. The budget that arrived last and got cut first.
That classification was always wrong, and AI is proving it. As the technical layer of work becomes automated, the human layer is what remains to distinguish one organization from another.
There is a second shift underneath the first, and it runs straight through your function. AI has made producing communication nearly free. Any organization can now generate more messages and updates than its people could ever absorb.
When production is free, the scarce resource becomes judgment. What deserves to be said at all. What this moment calls for. What a message will do to trust after people receive it. Whose voice should carry it.
Every employee with an AI tool can now produce polished communication. Very few can tell you whether it should exist.
That judgment is built the slow way, through decisions owned and consequences lived with, then strengthened by lessons closed out. Much of that building happened inside the early, effortful work we are now handing to machines.
Organizations that automate the work without asking where their people will build judgment are solving this quarter’s efficiency problem by creating the next decade’s leadership problem.
Communications and people leaders have more influence here than they may realize. Your function shapes how an organization speaks, how its leaders behave, and how its people grow.
The case you have spent years making—that trust and human connection are business infrastructure—is the case the entire enterprise now has to absorb. Make it openly. Ask the harder questions before anything ships. Teach your teams to form a view first, then use AI to pressure-test it instead of asking it to produce the view.
A generation of capable professionals was told that strength in people, communication, judgment, and leadership was a lesser form of ability. That was a costly error when I started my career, and it is about to carry a much higher price.
Organizations that hollow out their human capability while automating their technical capability will discover they kept the wrong thing.
We spent decades calling the hardest work soft. It took machines mastering the easy part to make the difference plain