Writing Better Prompts

What you'll learn

• Use roles and context to shape tone, depth and expertise
• Choose between zero-shot and few-shot prompting
• Write examples that guide the model without being copied blindly
• Request structured output: lists, tables, JSON and more
• Refine prompts iteratively and apply constraints deliberately

1. 01Shaping the request

   1. 1Not startedRole promptingGive the AI a role or perspective so it picks the right vocabulary, depth and priorities — and learn where roles help and where they don't.8 min
   2. 2Not startedContext and instructionsSeparate the background the AI needs from the task you want done, so it understands your situation and knows exactly what to do with it.9 min
   3. 3Not startedZero-shot promptingGet good results from instructions alone, with no examples — and recognise when a task needs more than a zero-shot prompt.7 min
   4. 4Not startedFew-shot promptingShow the AI a few worked examples so it can match a pattern, format or tone that's hard to describe in words.9 min

2. 02Shaping the response

   1. 5Not startedProviding examplesChoose examples that teach the right lesson: diverse, representative, well-ordered — and not so similar that the AI just copies them.10 min
   2. 6Not startedOutput formattingAsk for lists, tables, JSON or markdown — and control length — so AI output is ready to read, paste or process without cleanup.10 min
   3. 7Not startedIterative promptingTreat prompting as a conversation: review the output, diagnose what's off, and refine with targeted follow-ups instead of starting over.9 min
   4. 8Not startedUsing constraints effectivelyUse limits on length, scope, style and content to focus the AI — while avoiding conflicting, excessive or purely negative rules.10 min