I spent Tuesday morning in the ocean. Three organizations kept running while I was in the water, and nobody covered for me. The thing covering for me was a system I built, and it costs a few hundred dollars a month.
It was an ordinary Tuesday. I surfed at dawn. Later that morning I went to the gym, in the middle of what most people would call the workday. Somewhere in between, a consulting practice, a marketing agency, and the regional AI association I founded all kept moving without me sitting at a keyboard.
That sentence would have read as fantasy to me not long ago.
I used to spend my mornings the way a lot of leaders do. Buried in my inbox before I was fully awake, trading messages back and forth with people trying to find a meeting time, opening my calendar, moving something, closing it, opening it again an hour later. The work only I could do sat waiting behind the work that anyone, or anything, could have done. My best hours went to my worst tasks.
Underneath that ordinary Tuesday now is an operating system I built. Two agents do most of what I used to do by hand. One is my chief of staff. I call him Jarvis. The other is my assistant, and her name is Iris. They live on my phone, which means they come with me into the water, to the gym, into the car. When a request comes in while I'm driving, I send Jarvis a voice note at a red light. When I want tennis on the calendar Tuesday at nine, I tell him, and it's there.
The surprise was not the efficiency. It was that I got myself back.
When I describe this to other people, the response I hear most is a version of the same sentence.
I'm drowning. I wish I had an Iris.
I understand it completely, because I was drowning too. So my first instinct is always to reach for the hours, and there are a lot of them. I track how much time the system saves me every week, and the number lands between 25 and 40 hours depending on what I'm building. The scheduling alone gives me back two to three hours I used to lose to calendar tetris, sending emails back and forth to sync people up.
But the hours are not the part that surprised me.
The part that surprised me is that this is fun.
I have a strength training program I paid for, built specifically for women, and for months the logging was a chore I dreaded. Sets, reps, weights, all tracked by hand in a way that made me want to skip the tracking and eventually skip the workout. Now I tell Jarvis the numbers out loud in the middle of a set, and he keeps all of it. The system I built to run three companies also, quietly, made it easier to take care of my own body. I did not build it for that. I did not see it coming. It's one of the best things it does.
That was the moment the whole project changed shape for me. I had not built a machine that does my work. I'd built one that hands me my life back around the work.
Most leaders point AI at their team. I pointed it at myself first.
Here is the belief I want to push against, because I think it's the most expensive mistake a leader can make with this technology right now.
Most executives treat AI as something you deploy to the workforce. A tool you hand down so the team moves faster. A line item aimed at everyone below you on the org chart, and almost never at the person at the top.
I did the opposite. I built it for myself first.
I want to be precise about how it runs, because the way it runs is the reason I trust it. Iris does not act on her own. She drafts a reply or creates a calendar invite, and the moment she does, Jarvis pings me. I look at what she wrote, I confirm it's right, and I send it. There's still a small manual step, and I keep it there on purpose. The judgment stays with me. Anything that leaves my accounts leaves because I decided it should.
There are things I never hand off at all. I don't let AI run a sensitive client conversation. I don't let it make a commitment on my behalf. I don't let it touch personal information that has to stay secure. Those lines are not up for negotiation, and they're not a sign that I trust the system less than I should.
The boundaries are not the limit on the value. They are the reason the value is real.
A leader who hands the whole thing over and stops looking hasn't built leverage. They've built a liability they can't see. Because I stay in the loop on everything that carries my name or my word, I can move fast on everything else without a knot in my stomach. Since I built it, the reach-outs stopped slipping. People are constantly asking for my time, for a meeting, for something from me, and the things that used to fall through the cracks don't anymore. Jarvis catches them. I decide what to do with them.
A team of twenty for a few hundred dollars a month.
Let me put the economics plainly, because they're almost hard to believe out loud.
What this system does across three organizations would have taken a large team and serious money the old way. A real chief of staff. An executive assistant. An operations coordinator. The salaries and the overhead behind all of them. Call it a twenty person team and a million dollars a year, and you wouldn't be far off.
Mine runs for a couple hundred dollars a month.
It stays that cheap because I'm careful about which model does which job. Sensitive work runs locally. Everyday tasks go to a smaller, faster model. Only the heavy reasoning goes to a frontier model. I'm not paying premium prices to have the most expensive intelligence in the room fold my laundry. Maintaining and improving the whole thing costs me a couple of hours a week, and the time it hands back is roughly five times that.
Here's the part that matters most for the work itself. When I was buried in scheduling and inbox management, the things that actually slipped were my deliverables. The proposals. The client work. The thinking people were paying me for. I was so busy answering email that the real work waited. Now the email drafts itself and goes out with my sign off, and my attention goes back where it belongs. My work is better and it's faster, and that happened because I stopped spending my sharpest hours on the dullest tasks.
So the one move I'd ask any leader to make is this. Before you roll AI out to your team, before you write the policy or buy the licenses or book the training, build something small for yourself and live with it. The more a CEO uses this personally, the better they lead a company through it. You can't lead people into a change you've never made yourself. The executives who get this right will be the ones who used it first, felt the difference in their own week, and can speak about it from experience instead of from a slide.
I went surfing on a Tuesday, and the work got done.
I think a lot about what this technology is for. Not what it can do, which is close to anything now, but what it's actually for.
My answer has gotten simple. It's for giving people their time and their attention back, so they can spend both on the things only a human can do. Lead. Connect. Create. Be present with the person in front of them. Stand in the ocean for twenty minutes before the day starts and feel like the day is already handled.
I get more time to think now. More time in nature. More time being a person instead of a bottleneck.
The system runs the parts of my work that never needed me.
What I do with the hours it gives back is the whole point.
