I don't miss the meetings

For about six years, my calendar looked like a chessboard. Video call, video call, a gap, another video call. Most mornings started early with my director, who lived in India and was finishing his day while I was starting mine. Then the SEO managers. Then the creative team. Then, more often than I'd like to admit, a meeting to figure out why something from an earlier meeting hadn't happened.

From 2017 to 2023 I ran an online business that did a lot of what Slacked does now: content, websites and the operations behind them. People did the work, about fifty of them at the peak, spread across a handful of countries and time zones. They were good people. The meetings weren't their fault. They were mine, and they were the system.

Why the meetings existed

Looking back, nearly every call was the same three questions in different clothes. Where is this? Who has it? What's stopping it?

We had tools. We had project boards, shared docs and a chat app that never stopped pinging. None of them was the truth. The truth lived in people's heads, so the only way to get at it was to put everyone on a call and ask.

Every hand-off was one more place for the work to wear down.

That works with five people. With fifty, it becomes a job. A brief left me, went to the director, then to a manager, then to a writer eight or nine hours ahead. Each step retold it a little. By the time the draft came back, it answered a question I hadn't quite asked. So we'd meet about it.

People called this scaling. I never liked the word. In practice it meant more hand-offs, and every hand-off was one more place for the work to wear down.

What replaced them

AI made most of what that business sold close to free, and the business didn't survive it. When I started again, I rebuilt the same kind of work with two AI teammates. Claude runs the operations, Muse makes and checks the creative work, and I steer.

Here's the part I didn't expect. There are no meetings. Not fewer. Not one.

What replaced them is one written task list that everyone works from, me included. Every task has an owner, a due time and one line that says what done looks like. Claude picks up the next task, does it and leaves the proof: the live page, the log line, the file with today's date on it. Muse checks the work against written rules. I read the list over coffee and step in only where a person is actually needed, which mostly means money, sign-ins and anything legal.

The three questions still get asked. The list answers them instead of a call. Where is this? It's on the list. Who has it? The list says. What's stopping it? There's a line for that too, and if something waits too long, it gets flagged to me.

What it isn't

The difference is where the mistakes show up. In the old setup, a mistake hid in someone's head until the next meeting. Now it's written down, with a time stamp, where anyone can see it. That's uncomfortable at first. It's also the only reason the system gets better.

If you're moving from meetings to a list

You don't need an AI team to try this. If you run anything with other people, or even with one AI, these are the four things I'd copy:

• Write down what done looks like before the work starts. A link, a file or a number. If you can't name it, the task isn't ready.

• Ask for proof, not status. "Send me the live link" beats "how's it going?" every time.

• Keep one list. Not a board plus a doc plus a chat thread. One place that holds the truth.

• Save meetings for decisions only a person can make. If a call exists just to find out where something is, the list is broken, not the calendar.

I spent six years on those calls. I learned a lot on them, and I liked most of the people on the other end. But I don't miss the meetings. What I have now is quieter, and the work gets finished more often than it used to.