Stop just chatting with AI. Build agents that do the work.
A self-paced course written around your life and your apps: from better prompts to your own agent, loops that run themselves, and setups that fix their own mistakes.
Sample video for a fictional trainer, "Marcus". Your course is built from your own intake.
Free: your ladder placement + a ready-to-use Level 1 prompt. Full course $39 one-time. No coding needed.
Googling it vs. AI Playbook For Me
😵 Googling it
- 20 tabs and generic videos for someone else's setup
- Tools you don't use, steps that skip the hard part
- No idea whether you did it right
- Safety is an afterthought
- Stuck? Back to searching
🚀 AI Playbook For Me
- Your agent, built around your own apps and your own week
- Step-by-step builds with a checklist at every level
- Safety built in: drafts only, you approve sends and payments
- Prompt Gym: practice, don't just watch. Type real prompts, get scored and coached
- Stuck-help inside every lesson (tutor chat coming soon)
- Climb at your pace, starting at your level
The course map: 4 levels
Each level hands a bit more work to AI, and adds the safety habits that go with it.
Prompt engineering
Talk to AI well: give it your role and real context, say what good looks like, iterate, and ask it to flag doubts.
You build: a personal prompt kit for your own recurring tasks.
Safety habit: answers can sound confident and still be wrong. Check facts before you act.
Build your first agent
Connect AI to an app you already use, give it standing instructions, and let it take steps when you ask.
You build: a draft-only agent, like an inbox helper that leaves replies in your Drafts folder.
Safety habit: drafts only, smallest permissions, and you confirm before anything is sent, paid for, or deleted.
Loop engineering
AI that runs itself on a schedule or until a job is done, checks its own work, and keeps a log.
You build: a scheduled loop with a done condition, caps, a self-check, and a "needs me" list.
Safety habit: attempt and item limits, watch the cost, and start in report-only mode.
Self-healing
Your loop notices when something breaks, retries what's safe, falls back to a labeled option, and alerts you.
You build: error rules, a fallback, clear alerts, and a safe failure drill.
Safety habit: never auto-retry sending or paying. Logins and money always go to a human: you.
How it works
Fully self-paced. No classes, no waiting on anyone.
Tell us about you
A short intake: your job, routine, pain points, the apps you use, your goals, and how you use AI today.
Get your course
Your dashboard shows all 4 levels. Each lesson is written for you when you open it, starting at your level.
Build, test, pass
Do the hands-on build, tick off the checklist, and move up. Your progress is saved in your browser.
A peek inside
An illustrative snippet for a fictional project manager at Level 1.
Level 1 prompt, written for your pain point
Your intake said mornings start with a flooded inbox. So your first prompt sorts it, and in Level 2 that same prompt becomes your agent's instructions.
App screens change often, so builds describe the goal of each step, not just where to click.
Real results coming soon
One simple price
Pay once for the whole ladder.
- All 4 levels, personalized to your intake
- Your own prompts, agent, loop, and self-healing build
- Hands-on exercises and pass checklists
- Progress dashboard
- "Stuck?" tutor chat (coming soon)
Questions
Do I need to code?
No. Builds use apps you already have and no-code tools. Each step is in plain English.
Can I skip ahead?
Yes. You start at your current level, but all four levels are open. Each one builds on the one before, so earlier lessons help.
Is a person teaching it?
No. Your lessons are written by AI from your answers, using our instructions and safety rules. It's fully self-paced.
Will my agents send emails or spend money?
Not in our builds. Everything starts draft-only or report-only, and you confirm before anything is sent, paid for, or deleted.
What about custom GPTs?
We don't build the course around them. OpenAI has said it plans to retire them, so lessons are built on the Grok Bot desktop app: agents, scheduled routines, and connected apps like Gmail, Drive and Calendar, with a no-code fallback.
What if an app looks different from the lesson?
Screens change often. Each step explains its goal so you can find the closest match.
What if it doesn't work for me?
Finish Level 1 and not saving time? Full refund.
Is it a subscription?
No. It's $39, once.