I wasn't on a crusade against meetings. I just wanted to stop drowning in them. My calendar looked like a wall of colored blocks, and the actual work happened in the cracks between them, usually after everyone else had logged off.
Then I started using AI around my meetings — not to replace them, just to make them shorter and sharper. Within a couple of months, my meeting time was cut roughly in half, and the meetings I kept were better.
Nothing here is dramatic. It's a handful of small habits. But small habits stacked on a calendar add up to entire afternoons returned to you.
AI cut my meeting time in half by attacking meetings from three sides: it prepared them (so they started focused), it captured them (so I didn't waste a meeting just to take notes), and it replaced the ones that were really just updates (so they never got scheduled). The meetings that survived got shorter because the AI handled everything around them, leaving only the part that needed humans talking.
Once I looked closely, every meeting was doing three jobs at once, and AI could take two of them.
Job one: prep. Figuring out what we're even meeting about. Job two: the actual conversation. Humans deciding things together. Job three: capture. Writing down what was said and what to do next.
Jobs one and three don't need a room full of people. They were just stuck to the meeting out of habit. Peel them off, hand them to AI, and what's left — the real conversation — is short, because that's all a meeting was ever for. This sorting — automate the routine, protect the human part — is the core argument of the honest truth about AI productivity tools, and it's the same experiment I ran when I replaced my team meetings with AI for a month. Harvard Business Review has documented just how much of the average workweek disappears into low-value meetings, which is exactly the waste this attacks.
Most meetings start slow because the first ten minutes are everyone remembering what's going on. That's wasted, every time.
Now, before a meeting, I have an AI assistant pull together a tight brief: what this is about, the relevant background, the open questions, and the one decision we need to make. Everyone reads it in two minutes. The meeting starts already warmed up.
This alone cut a chunk off every call. A meeting that begins with "okay, why are we here" is a meeting that runs long. A meeting that begins with "we're here to decide X, here's the context" gets to the point and ends.
The format I use is dead simple: purpose, context, the decision, the options. If the AI can't fill in "the decision," I question whether the meeting should exist at all.
This was the big one. I used to half-listen in meetings because I was busy writing down what people said. Splitting my attention made the meeting longer and worse.
Now an AI assistant captures the meeting — the summary, the decisions, the action items, who owns what. I just participate. Fully present, not transcribing.
Two things happened. The meetings got shorter because nobody was slowing down for note-takers. And the follow-through got better because the action items were captured cleanly instead of half-remembered.
A meeting where someone is busy taking notes is a meeting where someone isn't really in the meeting.
The notes also meant fewer follow-up meetings. Half of "let's circle back" is really "I didn't write down what we agreed." Solid capture kills the circle-back.
The shortest meeting is the one that never gets scheduled.
A lot of my recurring meetings were pure status updates — people taking turns reporting what they'd done. That's not a conversation. It's a newsletter read aloud, slowly. So I moved those to written updates, and an AI assistant summarizes them into one digest.
No meeting needed. The information flows better, in less time, and nobody sits through eleven people reciting their week. This is where the biggest time savings came from — not shorter meetings, but deleted ones.
The test I use: "Does this meeting need people to talk to each other, or just report?" If it's reporting, it becomes text. If it's a real exchange, it stays a meeting.
The resistance to this is almost always cultural, not practical. People defend recurring meetings because cancelling one feels like saying the work in it doesn't matter, or that the people in it aren't important. That's the wrong frame. Killing a status meeting doesn't demote the work — it respects it enough to stop making everyone narrate it out loud. When I explained it that way, the pushback faded fast. Nobody actually enjoyed reciting their week to a screen full of half-listening faces. They just assumed the meeting was load-bearing because it had always been on the calendar. Most of them weren't. They were habit wearing the costume of importance.
Here's roughly how the time broke down, as my own real numbers, not a study.
| Source of savings | Roughly how much |
|---|---|
| Deleted status meetings | ~40% of the savings |
| Shorter meetings (better prep) | ~35% |
| Fewer follow-ups (clean notes) | ~25% |
The lesson hiding in that table: most of the win came from fewer and shorter, not from some clever trick inside the meetings. AI did the boring surrounding work, and the meetings shrank to their actual purpose.
One warning, because I nearly fell into it. It's tempting to automate so much that you delete meetings that should exist.
Some meetings are about trust, alignment, or simply being a team. Those aren't inefficiencies to optimize away. When I got too aggressive, the team felt disconnected, and I had to add a genuine human meeting back on purpose.
So the rule isn't "fewer meetings always." It's "fewer update meetings, protected human meetings, and AI handling the work around all of them." Cut the waste, keep the connection. AI is for the first part only.
Cutting meeting time in half sounds like a one-time win. It isn't. The real payoff compounds, and that's the part I underestimated going in.
The obvious gain is the hours back. But hours in a fragmented calendar are worth far less than hours in a block. Five separate one-hour gaps between meetings are nearly useless for deep work, because real focus needs a runway — time to load the problem into your head and stay there. A morning chopped into pieces by meetings produces almost no deep work, no matter how many total "free" minutes it contains. Cut the meetings and those fragments fuse into actual blocks, and the value of each recovered hour roughly doubles a second time.
Then there's the energy. Meetings, especially low-value ones, are quietly draining. You leave a pointless status meeting a little flatter than you went in. Stack five of those and by mid-afternoon you're running on fumes, and the work you do in that state is worse. Removing the junk meetings didn't just give me time — it gave me time when I was still sharp. That's a multiplier you can't see on a calendar.
It compounds socially, too. When you protect people's time, they notice, and they start protecting yours back. The culture shifts from "block an hour just in case" to "is this actually worth a meeting?" Once a team starts asking that question by default, the junk meetings stop getting scheduled in the first place. The savings become self-sustaining instead of something you have to keep fighting for.
So the honest framing isn't "I saved 50% of my meeting time." It's "I got back time, in usable blocks, while I still had energy, on a team that increasingly defends its own focus." Those four things stack. That's why a change that looks like a simple efficiency tweak ended up reshaping how much real work actually got done, not just how many meetings I sat in.
If your calendar feels like a wall of blocks, try the "talk or report?" test on next week's meetings and let AI handle the prep and capture on the ones that survive.
Q: Doesn't AI note-taking miss nuance? It catches decisions and action items reliably, which is what notes are actually for. The nuance lives in the conversation you're now fully present for, instead of half-missing while you scribble. Net, you lose less.
Q: Won't people just ignore the written updates? They will if the digest is long. Keep the AI summary short, blockers at the top, and people read it. A digest you can finish in under a minute gets read.
Q: How do I know which meetings to cut? Ask whether people need to talk to each other or just report. Reporting becomes text. Real exchange stays a meeting. That one question sorts almost everything.
Q: Isn't all this prep work just moving the time around? The prep is automated, so it's nearly free time, not your time. You're trading your expensive human hours in a long meeting for an AI assistant's instant brief. That's the whole point.
I didn't cut my meetings in half by being disciplined in them. I did it by letting AI handle the prep, the capture, and the updates that were never meetings in the first place. What's left is short, because what's left is the only part that ever needed a meeting.
Most of a meeting isn't the meeting. Automate the rest and watch it shrink.
Look at next week's calendar. How many of those blocks need people talking, and how many are just reports in disguise? Answering that honestly is the first hour you'll get back.
One person, output that looks like five. It isn't about working more hours — it's about a kind of leverage teams rarely have.
One idea a week to a published issue in under an hour. The boring system behind a newsletter I never dread sending.
Behind a lot of lean, profitable companies is the same small stack of AI tools. Here's what's actually running the show.
Comments
Sign in to join the conversation
No comments yet. Be the first to share your thoughts!