Most of the AI conversation is about what it produces. More output, more done in a day. I care more about what it does to the person using it. Is it helping me show up more fully for the people and the work, or quietly doing my thinking for me?
Underneath that question is a belief I try to live up to. Good work, good ideas, and change that actually lasts all come from the same place… people who feel they matter, thinking and talking together. Not from tools. Not from speed. From people, in connection. Technology’s only job is to help that happen. The moment it stands in for it, it’s working against the thing that made the work good.
That belief is why I work in one order. People first, then process, then technology. It’s how I lead, how I advise, how I build. AI didn’t change the order. If anything, it made me better at the parts that come before the tools.
People-first, for me, starts with curiosity. AI never replaced mine. It sharpened it. I ask more questions now, not fewer. I follow the thread further. When something doesn’t sit right, I can chase it down instead of letting it go. The tool didn’t hand me answers. It got me to better questions, and better questions were the real work all along.
So I don’t use AI to get answers. I use it to think. To pressure-test my own reasoning before I bring it to anyone. To fact-check myself before I check anyone else. To find the hole a smart skeptic would find, while there’s still time to fix it. Mostly, to learn out loud instead of pretending I already know. That last one matters more than it sounds. A lot of us were trained to project certainty. The most useful thing I’ve done with AI is get comfortable saying “I’m not sure yet, let’s look,” and then actually looking.
But the more I use it, the clearer one thing becomes. The best thinking still happens between people. Ideas rarely arrive fully formed in one head. They come from collision, from the conversation that takes a half-formed hunch somewhere you couldn’t take it alone. AI can be a good partner for that, but I’ve learned to be careful how I bring it in. A model is trained on the aggregate of everything already said, so it pulls, gently, toward the middle. And once it shows you an answer, you can’t un-see it; its framing quietly sets the options you think you’re choosing between. So I try to form my own read first, then bring it in to widen and pressure-test, not to seed. I ask it what I’m missing, not what to think. Treated as one more voice in the conversation, it makes me sharper. Allowed to become the conversation, it hands me answers with no one behind them.
Technology comes last, on purpose. The tool serves the people and the process, not the other way around. When technology leads, you get solutions looking for problems and platforms nobody uses. It’s the same reason change so often fails. A change lasts when the people it touches feel they matter in it. Noticed, brought in, needed. Then they own it, and what people own, they keep. Push it through process or tooling and people feel done-to. They comply while you’re watching and let it fade the moment you look away. Technology was never what makes change stick. People who feel ownership are.
Then there’s the quieter risk, the one I watch most closely. The parts of the work that matter most can’t be put into a prompt. Thirty years of pattern recognition, the read on a person, the sense that something’s off before I can say why, the ethical gut check, none of it fits into words cleanly, because we know more than we can say. A tool only works with what can be said. So on the calls that actually matter, it’s always missing the variable that decides them. That’s not a gap better models close. It’s where a person has to stand.
And it’s the same ground as making someone feel they matter. Being seen. The hard conversation, had honestly and made safe. The judgment call with your name on it. The thank-you or the apology that should come from you. AI can draft any of it, well enough that the temptation is real. But the second it stands in for the human part, it costs far more than the minutes it saved. Knowing what to keep human is its own skill, and I’m still learning where the line sits.
I’ll be honest about the rest too. AI is confidently wrong often enough that I verify before I trust it. It flattens your voice if you let it. It’s no help at all with the things that need a person to be present. None of that makes it less useful. It makes it a tool, which is exactly what it should be.
So here’s the test I hold every use to. Did this help me show up better for the people and the work, or did it just add to the noise? That question kills a lot of drafts. It slows me down before I hit send. And it keeps me honest, because “I could generate this” and “this is worth putting into the world” are not the same thing.
That’s what I want to share here. Not the tools. The thinking. Better questions, honest verification, and the parts worth keeping human. Some of it will be about where AI helps. Some about where it doesn’t. All of it starts with people.
So I’ll ask you what I keep asking myself. Is AI changing what you produce, or how present you are for the people around you?