A prompt is the input you provide to an AI system — a question, instruction, or piece of text that tells the system what you want. The quality of your prompt affects the usefulness of the output, though it cannot guarantee factual accuracy.
What Makes a Prompt Effective
Effective prompts tend to share several characteristics:
- Clear instructions. State what you want the system to do. “Summarize this text in three bullet points” is clearer than “what do you think of this?”
- Relevant context. Provide background the system needs to give a useful response. If you are asking about a specific policy, include the relevant details.
- Specified output format. If you need a list, a table, a paragraph, or a specific length, say so.
- Defined constraints. If there are things the response should or should not include, state them explicitly.
Practical Examples
Less effective: “Tell me about data governance.”
More effective: “Explain data governance in plain language for a municipal employee who has no technical background. Focus on why it matters and give two practical examples. Keep the response under 200 words.”
Less effective: “Fix this.”
More effective: “Review the following paragraph for clarity and grammar. Suggest specific changes. Do not change the meaning.”
Iterative Refinement
Prompting is often iterative. If the first response is not what you needed, refine your prompt rather than accepting an inadequate result. You might add more context, clarify the format, or narrow the scope.
Asking for Sources or Uncertainty
You can ask an AI system to indicate when it is uncertain or to note where a claim would need verification. This does not guarantee the system will accurately identify its own limitations, but it can prompt more cautious output.
Protecting Information
Do not include personal information, confidential organizational data, or sensitive details in prompts sent to external AI services. Treat prompts as potentially visible to the service provider. Related learning: Privacy 101.
Reviewing Generated Output
Always review AI-generated output before using it. Check for:
- Factual accuracy — verify claims against reliable sources
- Completeness — important information may be missing
- Tone and appropriateness — the output may not match your context
- Fabricated references — citations, names, and statistics may be invented
A well-written prompt improves the likelihood of a useful response. It does not make the response factually reliable. Verification remains the responsibility of the person using the output.
- Clear instructions, relevant context, and specified format improve prompt effectiveness
- Prompting is iterative — refine based on results
- Never include personal or confidential information in prompts to external services
- Always review and verify AI-generated output before use
- how to write AI prompts
- prompt writing tips
- AI prompt examples
- effective prompting techniques
- AI output verification