Skip to Main Content

Prompt Engineering

This guide provides a definition of prompt engineering, prompt engineering strategies such as the CLEAR framework and Zero-, One-, and Few-shots prompting, and other topics related to creating effective prompts for Gen AI tools.

Using the CLEAR Framework for Prompt Engineering

The CLEAR framework is an acronym that describes 5 main components or guidelines for constructing effective prompts. Watch (or listen to) the video, then review each step of CLEAR.

CLEAR Framework Steps Summarized

C - Concise

  • Use clear, simple language to instruct the AI model to get the results you want.
  • Prioritize important information.
  • Remove irrelevant words or details.
  • Skip polite words like please or thank you.

L - Logical

  • Write prompts so that they are coherent and follow a logical order.
  • Prompt for steps in a process.

E - Explicit

  • Provide specific instructions about output format, content, or scope.
  • Assign a persona or role like teacher, fictional character, etc.
  • Specify tone, reading level, or other specs.
  • Provide examples.

A - Adaptive

  • Be prepared to modify your prompts based on the results you are getting.
  • Rephrase, or try different approaches.
  • Split up question or tasks into smaller chunks.

R - Reflective

  • Evaluate generated AI output for accuracy.
  • Use Lateral Reading to fact-check.
  • Keep identifying how to craft better prompts.
  • Collect insights and learn from your experiences.