Generative AI: An Introduction

A prompt is the content you enter into a generative AI tool to query the application for some type of output.  This output might be text, image, or some other type of product.  The quality of the output you receive will depend in part on the quality of the prompt you provide.  Users often find they need to revise and refine their prompt to obtain the desired results.

Prompt engineering is the process of structuring text (or other input) in order to obtain quality results.  

Here are some tips for prompt engineering generative AI queries:

• Be clear and specific.  Long and confusing prompts can be confusing and may contain contradictory information that will confuse the generative AI application.  
• Provide context or give an example.  Give the generative AI tool some background information related to your question. One way to provide a context is to place the query in role.   
•  Break down complex queries into smaller parts and avoid "stacking" queries.  Often users want multiple pieces of information about a topic.  Breaking the query into parts and asking follow-up questions that add variables can improve the results.  Asking one prompt, then building on those results with another prompt is an excellent way to tailor your results.
• Include information about the type of output needed.  Depending on the tool, generative AI can provide code, narratives, stories, reports, audio, images, and much more.  Be specific about the type of results you need.
• Review your results and check for accuracy.  Most generative AI tools will include a warning that they are in an early stage of development and may return inaccurate information.  The results will sound authoritative but always verify important information (including citations to research)  independently of the generative AI tool.