October 2025
- Prompt Frameworks:
- POP – Persona, Objective, Parameters
- Task, Context, Examples, Persona, Format, Tone
- Be Specific: Detail exactly what you want. The more specific your question, the more targeted and relevant the tool’s response will be.
- Provide Context: If your question relies on specific context or background information, include that in the prompt. Context helps the model understand the scenario better and tailor its response accordingly.
- Use Clear Language: Avoid vague terms and jargon unless necessary. Clear and straightforward language helps ensure that the AI understands your prompts correctly.
- Frame Your Question Effectively: Think about what you’re asking and the best way to ask it to get a useful answer. Whether it’s open-ended, yes/no, or something that requires a detailed explanation, how you ask matters.
- Set Expectations: If you need information of a particular type (like a summary, a detailed explanation, bullet points, tailored to a certain audience, length of response etc.), specify this in your prompt.
- Break Down Complex Questions: If you have a complex question, break it down into smaller parts and ask them one at a time. This can help you get more comprehensive and precise answers.
- Use Follow-Up Questions: If the initial response isn’t quite what you were looking for, ask follow-up questions to narrow down or expand on the information.
- Check for Accuracy: While these tools aim to provide reliable information, it’s always a good idea to double-check its work.
- Learn from Responses: Pay attention to how your questions and the model’s answers correlate. Over time, you’ll get a better sense of how to phrase your prompts to get the best information.
- Provide feedback: It is likely the initial response won’t be exactly what you want, due in part to your prompt and its knowledge. Provide feedback, what was good and what needs to be improved and ask it to fix things.
- Ask for examples or clarification: If something is unclear don’t hesitate to ask how it got to that result or ask it to provide an example.
