🎓 delegation.school

Stop asking AI questions.
Start delegating to it.

The difference between a chatbot and a teammate is whether you ask it for advice or hand it real work. These lessons teach you to cross that line — one small, hands-on win at a time.

Start the first lesson →

Start from scratch

Never delegated to an AI agent before. Cross the line from asking to delegating.

  1. 1 The mindset flipWhy AI is a teammate with hands, not a smarter search box.
  2. 2 Your first real delegationHand off one real task and feel the agent act, not advise.
  3. 3 Start with a briefing, not a blank promptThe highest-leverage first prompt of your day isn't a task — it's "what should I do?"
  4. 4 Give it handsConnect one real tool and stop copy-pasting between AI and your work.
  5. 5 Make it remember youCorrect it once, tell it to remember, and it gets better every week.

Intermediate

You use AI daily. Connect it to your real tools and make it repeatable.

  1. 1 Connect your whole stackMove from one connected tool to the whole set, and run a task that spans all of them.
  2. 2 Stop starting from scratchTurn your most-repeated prompt into a saved workflow you run by name.
  3. 3 Give it a decision ruleHand it a standing ranking so it stops bouncing trivial choices back to you.
  4. 4 Second opinions on what's expensive to get wrongFor high-stakes work, get more than one model to weigh in and trust where they agree.

Power

Push it hard — subagents, custom skills, multi-model review.

  1. 1 Subagents & parallelismFan work out across parallel agents instead of waiting in a single thread.
  2. 2 Saved skills & slash commandsTurn a repeated procedure into a permanent, named capability.
  3. 3 MCP servers — give the agent new handsWrap a system you own as a verb-first tool the agent can call.
  4. 4 Multi-model verificationFan high-stakes decisions out to several models and act on the consensus.
  5. 5 Memory & self-improving loopsCapture what you learn so the agent compounds instead of resetting.