When I run AI workshops, I ask participants early on: “What’s the first agent you want to build?”
The answers are predictable. Email inbox management. Scheduling. CRM research. Lead enrichment. These are the obvious ones — the automations that sound impressive when you describe them at a dinner party.
Nobody ever says meeting prep.
And yet, after building agents for dozens of clients, the meeting prep agent is consistently the one that produces the “I can’t go back” reaction. Not the email agent. Not the one that does competitor research. The boring, invisible one that runs every morning before anyone else is awake.
What a Meeting Prep Agent Actually Does
The setup is straightforward.
Every morning at 7AM, a briefing lands in your inbox. It lists every meeting on your calendar for the day. For each person you’re seeing, it includes a brief summary — who they are, your last interaction, what the meeting is about — plus a direct link to their LinkedIn profile.
Thirty minutes before each call, you get a second nudge: a shorter message with the key context you need going in.
That’s it. Nothing flashy. But here’s what it replaces: 20-30 minutes of manual research before every single meeting. Digging through old emails to find the thread from six months ago. Opening three browser tabs to refresh yourself on who this person is. Glancing at their LinkedIn in the parking lot because you forgot where they work now.
The agent does all of it. You just show up.
Why This One Hits the Hardest
I use something I call the 80/20 rule for agent building: automate what you do every day, not what’s impressive.
Most people get this backwards. They want to start with the complicated research agent or the multi-step workflow that runs once a quarter. Those sound better at a dinner party. But they miss the compounding math.
Meetings happen every single day. If you have four calls per day and spend 20 minutes prepping for each, that’s 80 minutes of daily manual work. Multiply by five days a week. That’s almost seven hours a week of scrambling for context you could have had automatically.
An agent doing that in seconds isn’t a nice-to-have. It’s a different life.
The Discovery Pattern
What’s interesting isn’t just the time saved. It’s how people experience it.
Week one, the reaction is usually mild. “Oh, this is pretty helpful.” They appreciate it, but they’re not hooked yet.
Week two, something shifts. They start trusting it. They show up to calls more prepared than usual. They stop scrambling. And then at some point, they miss a meeting they forgot to check the agent for, and they feel the gap immediately.
“I don’t know how I ran meetings without this.”
That phrase — or some version of it — comes up over and over. It’s the discovery pattern of the best agents: you don’t know you want it until you see it, and once you have it, you can’t imagine doing without it.
Most tools ask users to tell them what they need. The best agents show users something they didn’t know to ask for, and then prove it in the first week.
Where to Start With Agent Building
If you’re trying to decide where to begin with AI agents, the framework is simple.
List the things you do every single day. Not the monthly reports. Not the once-a-quarter strategy review. Daily tasks. Recurring pain. The stuff that happens at 8AM before you’ve had your coffee.
Then ask: which of these is still done manually?
Meeting prep is usually near the top. So is inbox triage, daily briefings, and follow-up scheduling. Start there. Build the boring, daily, invisible thing first.
The impressive stuff can come later. But the compounding ROI of automating something you do every day will change your relationship with AI faster than anything else.
You don’t know you want it until you see it. And once you’ve seen it, there’s no going back.
P.S. Meeting agents, email agents, and the full AI workflow stack are part of what we cover in the Two Hour Workday program.
Recommended for you
Want the full system? 25X is the flagship productivity system we teach.
Explore 25X →Thanh Pham
Source link