Roles, Audience and Context

“Explain APIs” gets a different answer depending on who’s asking and who’s answering. Two quick additions control that: a role for the AI and an audience for the answer.

Give the AI a role#

Starting with “You are…” sets the expertise, vocabulary and point of view:

You are a senior Java developer reviewing a junior's code.
Point out bugs first, then style issues. Be kind but direct.

You are a patient school teacher. Explain fractions to a 10-year-old
using a pizza example.

You are an interviewer at a product company. Ask me one SQL interview
question at a time, wait for my answer, then give feedback and a score
out of 10.

The last one turns the AI into a mock interviewer: one of the most useful study tricks there is.

Name the audience#

The role decides who writes; the audience decides who reads:

Same topic: “What is cloud computing?”	Effect
“…for my grandmother”	everyday comparisons, no jargon
“…for a BCA student’s exam answer”	definitions, types, advantages, diagrams in words
“…for a CTO deciding between AWS and Azure”	costs, trade-offs, migration risks

Give it the background#

The AI only knows what’s in the conversation. Paste or describe what it needs:

Context: I run a small bakery in Surat. Most orders come on WhatsApp.
I want more customers on weekdays, when sales are 40% lower.
Task: Suggest 5 low-cost weekday promotions I can announce on WhatsApp
and Instagram. For each, write the message I'd post.

Without the context, you’d get generic marketing tips. With it, you get messages you can post today.

Build reusable templates#

When a prompt works well, keep it with blanks to fill in:

You are an expert {SUBJECT} tutor.
Explain {TOPIC} to a {LEVEL} learner who already knows {KNOWS}.
Use one real-life example and one short code/worked example.
End with 3 quiz questions (answers at the very end).

Fill it in: SUBJECT = Python, TOPIC = dictionaries, LEVEL = beginner, KNOWS = lists. You now have a personal tutor for any topic in 10 seconds.

Persona prompts for studying#

Act as a strict examiner. I'll explain how HashMap works in Java.
Find every gap or mistake in my explanation, then give me a model answer.

Explaining a topic and getting critiqued is one of the fastest ways to learn, and the AI never gets tired of listening.

Quick check

You want a beginner-friendly answer, but the AI keeps using heavy jargon. What do you add? An audience (“for someone with no technical background”) and a constraint (“explain any technical term the first time you use it”).

Summary#

• A role (“You are…”) sets expertise and tone.
• An audience (“…for a beginner”) sets depth and vocabulary.
• Give the AI the facts about your situation; it can’t guess them.
• Save prompts that work as templates.