Perfect prompting to generative AI programs
AI prompting, or, instructing AI to do a task, is a seemingly simple step to working with generative AI. But once you type your question and hit enter, has your first thought been, “That’s not what I wanted?” Building your prompting skill can improve the quality of your outputs now, and coach AI engines to be more accurate in the future.
Axios HQ offers ways to get more targeted results from your prompts to generative AI:
- Be specific with your ask and weed out the chances of inaccurate responses.
- Redefine your prompt with different verbs, questions, or identifiers for refined results.
- Use strong verbs.
- Ask AI to write its response in a tone suitable for your audience (formal, funny, casual, etc.).
- Define the length you want from AI’s response. You can ask for word count, page count, average read time, and more.
- Tell AI to format content in a specific way, such as a list or news release. For more visual formats like graphs and charts, AI can synthesize your data and give step-by-step instructions on creating them in a chosen software.
- Emulate your audience by denoting specific phrases or voice in your prompt that will generate relatable content.
Example prompts could look like this:
- “Write a playful but professional 200-word blog post about the benefits of communications best practices for college communicators.”
- “Create a bullet point outline of what should go on a back-to-school communications plan for a college biology department. Add a summary at the end.”
- “Write a funny, 50-word Instagram caption for a post asking for responses to the question, "What is the best food on campus?" The audience is Gen Z college students.”
Working with generative AI is all about experimenting. However you plan to use AI, remember to review its output, cite its use, and consult University Relations’ other resources for ways to use it in your role.