Nine short lessons that teach prompting the way you already delegate, with every move shown working. You see the naked request, what it actually returns, the built brief, and what changes. You leave holding five reusable briefs, a file of your own best work, a grounding prompt and a four-step sequence that makes AI argue with your decisions before anyone else does.
You type one line into the AI, get something vague back, and quietly conclude the tool is overrated. It is not. You gave it an instruction you would never accept from a person, and it did what a person would have done with it, which is guess.
A prompt is a delegation brief. That is the whole reframe, and it works because you already brief people for a living. What is missing is repetition against real work with the output in front of you, which is what this course is.
Every move here is run rather than described. You see the request most executives send, the output it produces, the built version, and the difference between them. Where the point is quality, the output is on the page.
CEOs and owners who have tried these tools, been underwhelmed, and suspect the problem might be how they are asking.
Executives and functional leaders who want dependable output for real work: board updates, analysis, difficult communication, decisions under pressure.
Anyone who has watched a colleague get sharp results from the same tool that hands them mush and wants to know what the colleague is doing differently.
Expected time: about 34 minutes across nine short lessons, plus the ten minutes it takes to assemble the two files the course asks you to build.
The new-hire test, which catches most bad results before you waste a round on them.
The five-part brief, and the reason it gets faster rather than heavier as you use it.
Three role patterns including the adversarial one, and the honest limit: a persona changes emphasis, never knowledge.
A three-line context block about your role and preferences that you write once and paste forever.
Your gold-standard file: five of your own artifacts that show the model what good looks like, labelled by what each demonstrates.
Four output shapes worth memorising, and why asking for named columns does part of the analysis for you.
The "without" habit: one explicit condition on every prompt you would sign your name to.
A copy-paste grounding prompt that stops invention, plus a two-line paste rule you can hand your team.
A five-step pre-mortem that turns the model into the sceptic in the room, and the trap of treating its critique as a verdict.
A maintained five-entry prompt library, seeded with briefs you watched get built.
Prompt outputs shown throughout are shortened illustrations of what each prompt produces, not transcripts of a particular session.