Replay · Tuesday Live Training

Agents vs Chatbots. Stop getting to-do lists. Start getting work done.

Most owners use AI like a chatbot: ask a question, get advice, and walk away with more work. An agent runs on a schedule, sees your real data, does the task, and reports back. Listen to the full training: the four-question test, the hand-off grid, and the prompt that finds your first task to hand off.
November 10, 202667 minutes
What's Covered

The full framework.

04:53
One prompt, a whole life plan.
The Flight Plan prompt: a 20-year North Star broken down to this month, a daily march, and a dashboard that tracks it. It asks the questions you weren't thinking about.
12:17
If I had to remember to open it, I wouldn't.
The dashboard that shows up every morning, the habits it puts on the calendar, and the end-of-day check-in that rewrites tomorrow. Proactive, not waiting to be asked.
19:02
Which one is your week?
Nothing moves unless you touch it. First in, last out, still behind. Or AI made you faster at work you shouldn't be doing at all.
20:13
My dad did what was available.
Building a global business used to mean a new city every week and missed games at home. There's a better way now, and it runs from the living room.
24:20
Chatbot vs. agent.
A chatbot starts when you ask and ends with you copying and pasting. An agent starts on a schedule or a trigger. When X happens, this SOP runs. Plus the content gold agent that mines every class.
30:53
Real systems, not guesses.
Why AI makes things up, and how connecting it to the source (CRM, inbox, calendar, shared drive) fixes it with a clickable reference. When private data calls for a small local model instead.
36:11
The four-question test.
Does it see your real data? Does it take action? Does it run without you? Does it report back? Four yeses, or it's a chatbot with a nice name.
51:29
Your first hand-off.
The often-or-rare, rules-or-judgment grid. Keep it, have it drafted, write the checklist, or hand it off. Then the calendar audit and the take-home prompt that finds your first task.
Key Takeaways

What you'll walk away with.

01
The four-question test.
Sees real data, takes action, runs without you, reports back. Run it on every AI tool you use today and you'll know which ones are agents.
02
Triggers and schedules.
Every agent starts with a moment: a new client, a finished class, every Monday at nine. Name the event, name the SOP, and the work starts without you.
03
The hand-off grid.
Rare and needs judgment stays with you. Often and needs judgment gets drafted. Rare and follows rules gets a checklist. Often and follows rules comes off your plate first.
04
Migrate once.
Teach one brain who you are, what you do, and how you do it. When a new AI shows up, plug it in by MCP instead of rebuilding everything from scratch.
Take-Home Prompt

Find your first hand-off in ten minutes.

Paste this into ChatGPT, Claude, or any AI you use. Turn on voice mode and talk it through, and give it your calendar if you can. It finds the task. It doesn't do the task.

Interview me about my week, one question at a time. Ask what I did each day, roughly how long each task took, and which tasks repeat every week. Then sort my tasks into four groups: often and follows rules, often and needs my judgment, rare and follows rules, rare and needs my judgment. For the "often and follows rules" group, rank the top three by hours per month, and for each one tell me what an agent would need to see, do, run on, and report back.

Ready for the Next Step?

The prompt finds the task. An agent takes it off your plate.

Executive Office AI gives you the agents, the prompts, and a live session every Tuesday to build them on your business.

Next Tuesday live: The Year-End Sprint →