Anatomy of a good prompt

The quality of your prompt determines the quality of the output. A good prompt has a clear structure — and learning that structure lets you get great results reliably. Let's dissect what makes a prompt good.

Vague vs good prompts

❌ VAGUE:
"Write something about fitness."

✅ GOOD:
"Write a 100-word motivational intro for a blog post aimed at busy
beginners, explaining why just 20 minutes of exercise a day makes a real
difference. Encouraging, friendly tone."

The good prompt says WHAT, FOR WHOM, HOW LONG, and IN WHAT TONE.

Compare a vague prompt ("write something about fitness") with a good one (specifying the task, audience, length, angle, and tone). The vague one leaves the AI guessing (generic result); the good one directs it (targeted, useful result). A good prompt tells the AI exactly what you want. The single biggest factor in output quality is how clearly and specifically you ask. AI can't read your mind — the more you specify, the better the result.

The parts of a good prompt

A good prompt often includes:

TASK:     what you want done (write, explain, summarize, list...)
CONTEXT:  the background (who it's for, your situation)
SPECIFICS: length, format, focus, tone
(optional) ROLE: "act as a [expert]"
(optional) EXAMPLES: show what you want

Example: "Act as a friendly career coach [ROLE]. Give me 5 tips [TASK]
for a nervous first-time job interviewer [CONTEXT], each with a specific
action to take [SPECIFICS]."

A good prompt typically has: a clear task (what to do), context (background — who it's for, your situation), specifics (length, format, focus, tone), and optionally a role ("act as...") or examples. You don't need all every time — but the more relevant detail, the better. Think of it as briefing an assistant: a good brief gets good work. These components (task + context + specifics) are the anatomy of a strong prompt. Include the ones relevant to your task, and the output improves dramatically.

Build prompts by adding detail

Start with the task, then add what helps: context, specifics, tone. Each addition makes the prompt more precise and the output better. You can build up a strong prompt this way — begin simple, then ask "what would help the AI do this well?" and add it. Over time, including the right details becomes automatic. But the habit — being clear and specific — is what reliably produces great results.

The mistake beginners make

Giving vague, minimal prompts and expecting great results (garbage in, garbage out). Also, withholding context the AI needs. And giving up after one vague attempt. Use the anatomy of a good prompt — task + context + specifics (plus optional role/examples) — including what's relevant. Being clear and specific is the #1 prompting habit. A good prompt is a good brief: the clearer you are, the better the output.

Your turn

Turn a vague prompt into a good one by adding the parts:

Vague: "Help me write an email."

Good:  "Write a polite email to my landlord requesting a repair for a
leaking kitchen tap. Mention it's been leaking for 3 days and ask for it
to be fixed this week. Keep it under 100 words, friendly but clear."

Notice how much better the result is with task + context + specifics.

---

Your turn

1. Understand the prompt determines output quality — vague prompts get generic results, good prompts get useful ones.
2. Learn the anatomy: task (what) + context (background) + specifics (length/format/tone), plus optional role/examples.
3. Build prompts by adding relevant detail — being clear and specific is the #1 prompting habit.

Key points

• The quality of your prompt determines the quality of the output — vague prompts leave the AI guessing (generic); good prompts direct it (useful).
• A good prompt has a clear TASK + CONTEXT (background) + SPECIFICS (length/format/tone), plus optional role or examples — like a good brief to an assistant.
• Include the components relevant to your task (more for complex, fewer for simple); build up by adding what helps.
• Being clear and specific is the #1 prompting habit — AI can't read your mind, so the more you specify, the better the result.