Writing Effective Prompts
A prompt is the instruction you give to an AI tool. The clearer and more specific your prompt, the better the AI’s response.
Prompting well is a skill — like searching effectively in Google or Library Search. The more you practise, the better results you’ll get.
Key Terms
- Prompt: Your question, instruction, or request to an AI tool. Can be text, images, audio, or files.
- Output: The result the AI produces (text, image, audio, or a mix).
- Prompt engineering: The process of crafting and refining prompts to get relevant, high-quality results.
What Makes a Good Prompt?
A good prompt is:
- Clear: Avoids vague or ambiguous wording.
- Specific: Includes details, context, and any constraints (e.g., word limit, tone, or style).
- Focused: Asks for exactly what you need.
- Iterative: Can be improved over several rounds.
Elements of an Effective Prompt
- Task: What you want the AI to do.
- Details: Key facts, requirements, or examples.
- Context: The background or audience.
- Boundaries: Word count, style, format, or limits.
Prompting for Different Content Types
Text
- Be clear and specific.
- Set tone or style (e.g., “formal academic tone” or “friendly explainer”).
- Add constraints like word limits or required structure.
Images
- Describe the scene in detail.
- Include style (e.g., “watercolour illustration” or “realistic photograph”).
- Use input images if the tool allows, then refine.
Audio
- Describe the soundscape or mood.
- Specify instruments, tempo, genre, or length.
- Give feedback to refine the result.
Prompt Frameworks
These can help you structure effective prompts.
PACE — Good for academic or research tasks
- Purpose — What do you want to achieve?
- Audience — Who is it for?
- Context — Background information needed.
- Expectations — Format, style, or boundaries.
CLEAR — Good for writing and explanation
- Context — Background.
- Limits — Length, scope.
- Examples — Style or sample.
- Action — Task to perform.
- Refine — Follow-up for improvement.
RACE — Good for problem-solving
- Role — AI’s role (e.g., tutor, editor).
- Action — Task.
- Constraints — Boundaries.
- Extras — Additional detail.