How to Use AI to Run Better Meetings

Meetings are the biggest, least-measured expense at any company. Add up the hourly cost of everyone sitting in the room, and you'll see that a weekly meeting with six people costs, over a year, roughly what one employee costs. And yet most meetings end the same way: a lot got said, something got agreed on that nobody wrote down properly, and the following week the same thing gets discussed all over again. AI isn't going to fix a bad meeting culture, but it does remove almost all the friction that makes meetings unproductive.

Before: Show Up Prepared in Five Minutes

A good chunk of a meeting is spent getting everyone up to speed: recalling what was said last time, summarizing where a topic stands, repeating numbers that already existed somewhere. All of that can be prepared beforehand in minutes. Asking AI for a summary of previous decisions, a list of open items, and a draft agenda with only the topics that actually require a decision, completely changes how the meeting starts. You walk in ready to decide, not to catch up.

During: Someone Takes Notes Without Anyone Having to Take Notes

The classic problem is that whoever takes notes doesn't participate, and whoever participates doesn't take notes. With a transcript and an automatic summary, that trade-off disappears: everyone participates, and at the end there's a faithful record of what was said. But what's genuinely useful isn't the full transcript, which nobody reads — it's the summary that matters: what got decided, what's still pending, and who committed to what and by when.

• Before: agenda, context, and open items from the last meeting.
• During: a record of what is said without sacrificing a participant.
• After: a summary with decisions, owners, and dates.
• Follow-up: automatic reminders for each commitment.

A meeting with no written decisions and no owner isn't a meeting. It's an expensive conversation.

After: Where the Real Payoff Happens

This is the point that turns a meeting into results. Most agreements don't fall apart out of bad faith — they fall apart because they fade: nobody remembers exactly what was said or who committed to doing it. If a summary comes out automatically at the end with each commitment, its owner, and its date, and those commitments trigger reminders, follow-through goes up immediately. It doesn't take more pressure or more follow-up meetings — it takes making sure what was agreed doesn't get forgotten.

The Rule You Can't Skip

If you're going to record or transcribe a meeting, tell the participants and get their agreement. It's not a legal technicality — it's a matter of trust, and a meeting where people suspect they're being recorded secretly is a meeting where nobody says what they actually think. Just say it up front and be clear about what happens to the recording. And if the topic is sensitive or confidential, simply don't record it and take notes by hand, like always.

Fewer Meetings, Not Just Better Meetings

There's a side effect worth more than everything above. When summaries and status updates get generated and shared on their own, a lot of meetings stop having a reason to exist: the weekly status meeting existed to keep people informed, and if everyone's already informed, it's no longer needed. Replacing two informational meetings a month with an automatic summary gives whole hours back to the entire team. That's the best use of AI here: not making meetings more efficient, but eliminating the ones that never should have existed in the first place.

Start with your heaviest meeting — the one with the most people and the least to show for it. Prepare it with a summary beforehand, close it with written decisions and owners, and see how it feels in two weeks. It's one of the simplest, highest-impact changes a team can make, and it doesn't require anyone to learn anything complicated. If you want your team to get the most out of everyday uses like this, it's exactly what we work on in our AI courses.