How AI Runs Our Meetings: Before, During, and After

Our team meets every day for a fifteen-minute huddle. It’s the group that builds the AI operating system we run our whole company on, and that daily meeting is now the lightest, fastest part of their day.

For most businesses, a daily team meeting like this runs thirty minutes or more and leaves everyone a little drained. Keeping a remote team in sync takes a relentless amount of coordination, and most of it happens in and around meetings: the status updates, the who’s-doing-what, the catching everyone back up. Day after day, that’s exhausting.

We’re not exhausted by meetings anymore. They’re actually pretty fun now. The change wasn’t a better meeting tool. We run our whole company on an AI operating system that already does a huge share of the work across the business. It’s also on our meetings, helping to prepare for and support the meeting as it happens, and execute next steps after.

Your AI notetaker is not the solution

By now, almost every team has an AI notetaker, or has at least sat on a call with one. Fathom, Granola, Otter, Fireflies. They join the call, record it, and hand you a transcript and a summary afterward.

But they don’t really solve the meeting problem:

• Multiple different notetaker bots sit silently in the same meeting, because everyone brought their own.
• A faceless “AI Notetaker” looms in the participant list, making people a little more guarded about what they actually say.
• Transcripts and summaries pile up afterward in a transcript app nobody actually opens.

Recording and transcribing a meeting is the easiest, least valuable thing AI can do for you.

You want execution, not notes

The impact of a meeting isn’t just the transcript. It’s the alignment that comes out of preparing for it, running it, and effectively kicking off the next actions. But that’s also where meetings get bogged down. Someone has to assemble the context beforehand so the conversation is sharp, and someone has to turn what got decided into assigned work afterward. Most teams do the first part badly and the second part late, if at all.

A notetaker can’t close that gap:

• It has no context beyond the meeting itself. It only knows what was said in the room.
• It doesn’t know your company’s strategy, your current targets, what got decided last week, or how the thing you’re discussing connects to everything else moving through the business.
• It has no access to the systems where the real work lives: your project management system, your docs, your credentials, your CRM.
• It can’t create tasks and assign them. And it definitely can’t execute them for you.

Strip it down, and a notetaker does just one thing: it notes the words of the meeting. You may not need to remember every detail anymore, but remembering words was never the point. The work output of a meeting was always the alignment and the next actions it sets in motion.

AI Runs Our Meetings (including next actions)

The AI in our meetings isn’t a bolt-on notetaker. It’s a system that’s layered across our whole business.

• Our work lives in Asana, our calls are on Zoom, and we work in other systems as needed.
• The AI is connected to all of them, the meeting transcript is just a record of the words of a meeting.
• We collaborate with the AI on meetings just like a human - and it handles next steps similarly too.

Here’s what happens before the meeting:

About ten minutes before the huddle, the meeting leader asks the AI to pull the full context and build the agenda. It posts straight to the Asana task for that meeting, organized into three to five sections:

• Progress against our current targets,
• What moved across the team since the last meeting,
• The blockers we need to remove today,
• and links out to the specific tasks for additional detail.

The AI takes three or four minutes to sweep the entire system and assemble that. Pulling the same picture together by hand would take a person hours, so honestly it just wouldn’t be done.

During the meeting, the AI helps manage:

We browse the agenda for two or three minutes. Then we spend the rest of the time on one thing: blockers.

• The AI knows the full context of what’s been done.
• If we have questions about specific details, we can ask the AI instantly.
• The AI understands dependencies between work, so we can know what in flight work will finish soon and how that unlocks the next work to be done.

For us, we chat with it using voice text during the call, on screen. This is faster than voice for us, but voice could work as well.

After the meeting, AI ensures everything is kicked off clearly:

This is the part that has historically been exhausting - tracking next steps, kicking them off / assigning them, etc. For us, it takes about 5 minutes.

When we wrap, the leader voice-texts a few sentences of direction. An actual example from our team:

“We unblocked the client email account contingency today, and the credentials will be in LastPass by 3pm PST. Make sure there’s a next action assigned to Mark to use those credentials to connect the client’s email to our system by end of day.”

Because the AI already has the meeting’s agenda task and the recording of the call that just ended, it understands that direction instantly. It also has access to where that work needs to happen (i.e. Asana) - so it can kick off next steps.

The notes get updated with what we discussed. The next action gets created and assigned to Mark in Asana. If it hasn’t progressed by tomorrow, it surfaces on the next agenda on its own.

It all just gets done (by the AI).

Now, meetings are fun again

This unlocked efficiency, bringing the huddle from thirty-plus minutes to under fifteen. And it didn’t only get shorter. It got better. We unblock problems now at a depth we never reached back when half the meeting went to status updates.

The time saved is the obvious win. But the bigger win is how it made our meeting routine fun again.

• Running a large remote team takes an enormous amount of coordination, and most of it lands in and around meetings.
• The AI absorbed all of that overhead.
• What’s left is the part that was always the point: a few people solving the one or two real problems of the day, together.

It brought us back to the fun part of collaborating. I would describe it as “delightful,” as cliché as that sounds.

It’s Not a Notetaker - It’s the Whole Business

A meeting recorder is a tool that creates a record of a meeting. It can’t prepare the meeting, answer questions inside one, or carry the decisions into real, assigned work.

But that’s straightforward for an AI that already knows your whole business. That’s how we operate, and we love comparing notes with other owners figuring out the same thing.

If you want to think through where AI fits in how your business runs, bring it to the free AI Strategy Brainstorm — an open discussion, not a pitch.