Long story cut short, we should make it a default practice to use the CREATE formula for prompts, so that we receive accurate, relevant, results those are aligned with your original goals. Now to understand CREATE, here you go :
C – Character Specify the role or persona you want the AI to adopt. For example, instructing the AI to behave as “a seasoned financial advisor,” “a thoughtful mentor,” or “a creative storyteller” provides a framework for tone and perspective.
R – Request Clearly articulate the core task or question. Describe exactly what you’re looking to get as a result—for instance, “summarize this report,” “draft an email invitation,” or “develop a marketing plan.”
E – Examples Provide one or more examples to illustrate the kind of response you expect. Examples can include sample phrases, formatting styles, or even a brief sketch of the answer. This helps guide the AI by showing it an ideal output structure or tone. For instance, if you want a list of bullet points, showing one or two example points can be very useful.
A – Adjustments Include any specific modifications you desire regarding style, tone, or detail. This might involve setting constraints (such as “avoid technical jargon,” “use a formal tone,” or “keep it under 150 words”) or asking for additional elements like a summary or actionable recommendations.
T – Type (Template) Specify the format or structure in which you want the answer. Whether it’s a step-by-step guide, bullet points, a table, a narrative, or even a code snippet—the clearer you are about the format, the better the final output will align with your needs.
E – Evaluation Instruct the AI to review or “evaluate” its output against your original goals. This step can include asking it to verify that all parts of the prompt have been addressed or to provide a short self-assessment of how well the answer meets your specifications. For example, you might say, “After generating the response, please assess if it meets the criteria mentioned above and note any areas of uncertainty.”
By using the CREATE formula in this way, you guide the AI through a focused process—from setting the persona to providing examples and requesting a review of the final answer—thereby enhancing clarity, relevance, and overall quality. This framework is particularly useful when crafting complex prompts where precise output format and tone are important.
PMI Team Member
