Prompting Frameworks
A prompting framework is a structured methodology or template for crafting effective and consistent prompts for AI models. These frameworks provide a systematic approach to defining the necessary components—like context, instructions, and desired format—to guide the AI in generating accurate, relevant, and reliable responses, rather than relying on trial-and-error.
Example Frameworks
Below are two frameworks that use mnemonic devices to help you remember key prompting elements and structure initial prompts effectively. Not only are these great ways to craft strong first prompts, they are also excellent tools for teaching students and others the basics of good prompting.
CRAFT – A framework for crafting clear AI prompts using Context, Role, Audience, Format, and Tone. Ideal for ensuring AI responses are relevant, understandable, and audience-appropriate.
PREP and EDIT – Use the PREP framework (Prompt, Role, Explicit, Parameters) to structure prompts that guide AI in completing specific tasks or adopting particular styles. This method works best for single-task prompts that need clear direction and boundaries. Then, apply the EDIT model (Evaluate, Determine, Identify, Transform) to review and refine the AI’s output for improved clarity, accuracy, and tone.
Frameworks are starting tools that prime the AI for optimized discussion. They provide a strong foundation so that your first prompt already produces a useful response, leaving iterative refinement and tweaks to take it the rest of the way.