Context Is Everything
The model can’t read your mind. Most “bad AI output” is really missing context.
You ask an AI to “write an email rescheduling my meeting,” and it produces a generic, robotic note you’d never send — wrong tone, no names, no actual new time. The instinct is “the AI is dumb.” The real issue: it had none of the context only you know. Your job this lesson is to take a vague request and make it land — by supplying context, not by hoping for better luck.
The model only knows what’s in front of it
An LLM has no access to your situation, your relationships, your goals, or your files unless you put them in the prompt. It fills the gaps with the most generic plausible answer — which is exactly why the output feels bland. The fix is almost never “better wording.” It’s more of the right context: who, what, the specifics, the tone, the constraints.
Context = who, what, specifics, tone, format
A rich prompt answers the questions a competent human assistant would ask before doing the task:
- Who is it for / from? (the client Dana; you, apologetic)
- What exactly needs to happen? (move Thursday 3pm to early next week)
- Specifics the model can’t guess (offer two alternative times)
- Tone and format (warm but professional; a short email)
Supply those and a capable model produces something you could send with little or no editing.
Vague: “Write an email rescheduling my meeting.” Rich: “Write a short, warm-but-professional email to my client Dana apologizing for moving our Thursday 3pm call. Propose Monday or Tuesday next week, morning preferred, and thank her for her flexibility.” → A model can nail the second one because nothing important is left to guesswork.
Briefing a new assistant. If you say “handle my email,” you’ll get chaos — not because they’re incompetent, but because they don’t know your world. Give them the who, the what, the tone, and the specifics, and they do beautifully. The model is the same: it’s only as good as the briefing.
Rescuing a vague request, step by step: 1. Vague: “reschedule my meeting.” The model invents generic details → unusable. 2. Add WHO: client named Dana; sender is apologetic. 3. Add WHAT + specifics: move Thursday 3pm; offer Mon/Tue next week. 4. Add TONE + FORMAT: warm-but-professional; short email. 5. Result: a near-send-ready draft. You changed the context, not the model.
This is the reading. The interactive version — active-recall quiz, a hands-on experiment you run in your own AI, and an earned mastery check — is free in the app.
Start this lesson free →