Seminar notes · Day 1
Your First AI Employee in 60 Minutes
Launch one supervised AI employee with a job description, tool access, an approval rule and a morning brief.
Ian Kirk did not invent these frameworks — they came from the operators teaching them, got tested against real work, and are taught here the way they actually get used. Every number belonging to somebody else stays marked as reported and second-hand.
Two questions sort a room fast. Who is using Claude or ChatGPT. Then, who is running agentic AI. Almost everybody puts a hand up for the first one. Almost nobody keeps it up for the second, and the gap between those two hands is the whole lesson.
Here is the difference. Agentic AI is working for you around the clock. You activate it whenever you want, or it activates on its own, for your chief of staff, your sales, your fulfillment, your operations. That is a different thing from a chat window you open when you have a question, and it needs different plumbing underneath it.
What follows is the build order. Hands first, then a runtime, then data and memory, then one job description, then one approval rule, then a morning brief. There is a specific product named at every step and every one of them will change. Claude will ship something, ChatGPT will ship something. Treat the product names as a snapshot and the sequence as the part worth keeping.
Free. One email, then the whole thing.
The rest is locked
8 sections, 5 frameworks, 6 steps to run. Free, one email.