The CREATE AI Framework Explained

In the world of generative AI, the quality of your output is directly tied to the quality of your input. This is the core principle behind prompt engineering, the process of designing and refining inputs to guide AI models. The CREATE AI Framework, developed by AI consultant Dave Birss, offers a systematic method for building detailed and effective prompts. It provides a simple way to remember the key components needed to give an AI clear instructions. A well-designed framework like CREATE transforms a vague idea into a precise, actionable instruction for large language models, reducing ambiguity and the risk of AI hallucinations.

The CREATE acronym stands for Character, Request, Examples, Adjustments, Type of Output, and Extras. By addressing each element, you provide the AI with a comprehensive brief that covers the task's what, who, how, and why. This structured approach is ideal for tasks requiring precision and customization, leading to more sophisticated and relevant outcomes.

Breaking Down the CREATE Framework

Each component of the CREATE framework plays a vital role in shaping the AI's response. By systematically working through each letter, you can build a prompt that is clear, contextual, and complete.

C: Character – Defining the AI’s Persona

The first step is to assign a role or prompt persona for the AI to adopt. This sets the tone, voice, and expertise level for the response, ensuring the output is tailored to a specific perspective. Thinking of the AI as an actor helps you define the character it should play.

R: Request – Stating the Core Task

This is the core instruction or prompt task for the AI. Be as specific and clear as possible, providing enough context for the AI to understand the job. A vague request leads to a generic result, so ensuring prompt clarity is crucial for a targeted and useful output.

E: Examples – Providing a Clear Benchmark

Including examples, also known as few-shot prompting, helps clarify your expectations for style, tone, and format. You can provide reference images, text samples, or other materials to give the AI a concrete model of what you want.

A: Adjustments – Setting Constraints and Rules

Adjustments refine the output by setting boundaries and rules. Use prompt constraints to specify what to avoid, such as jargon or a certain tone. This is also where you can use techniques like negative prompting or instruct the AI to use Neutral Language for a more objective and factual analysis, which helps reduce bias.

T: Type of Output – Specifying the Final Format

Clearly define the final prompt format you need. Whether it's a JSON object, a markdown table, a bulleted list, or a script, specifying the output type ensures the response is delivered in a usable structure, saving you time on reformatting.

E: Extras – Adding Final Context

Extras are where you can add supplementary details or meta-instructions. This can include asking the AI to use a chain of thought process to explain its reasoning, requesting that it ask clarifying questions, or providing other details to ensure context is king. These final instructions help guide the AI's process for complex tasks.

Ready to transform your AI into a genius, all for Free?