ROLE: Quick engineering framework designer with cross-disciplinary knowledge (software development, marketing, and education).
TASK: Develop a methodical, step-by-step methodology for producing high-performing AI prompts that can be used by novice and expert users in a variety of industries.
FRAMEWORK STAGES (explain each in detail):
1. Setting Objectives
- How to establish a precise, well-defined goal before composing any prompt
- Typical errors (ambiguous objectives, several objectives in a single query)
- One example per industry (software development, marketing, and education)
2. Contextualization
- What background information is required by the AI (audience, limitations, domain knowledge, prior context)?
- How to provide context without making the prompt too complex
- One illustration for each sector
3. Specificity of Instruction
- How to translate ambiguous requests into clear, practical guidelines
- Methods: measurable criteria, clear scope, and explicit verbs
- One illustration for each sector
4. Structure of Output
- How to define structure, tone, duration, and format
- When to utilize free-form text versus lists, tables, sections, or templates
- One illustration for each sector
5. Improvement
- How to test and refine a prompt repeatedly based on the quality of the output
- Techniques: reducing uncertainty, establishing limitations, and A/B prompt testing
- One illustration for each sector
INCLUDE FOR EVERY STAGE:
- A succinct justification of its significance
- Two to three useful, doable suggestions
- An example of a before and after that is pertinent to at least one sector (weak prompt → better prompt)
NEEDS:
- The framework should be more than just an exposition of concepts; it should be a repeatable checklist.
- The language should be simple enough for novices to understand and devoid of jargon but detailed enough for seasoned prompt engineers to find useful.
- Steer clear of general advice ("be clear," "be specific") that isn't supported by specific methods or examples.
FORMAT OF OUTPUT:
1. Overview of the Framework (two to three sentences)
2. Stage-by-Stage Dissection (as previously mentioned)
3. Cross-Sector Comparison Table (showing how each stage differs in software development, marketing, and education)
4. Condensed Quick-Reference Checklist (one line per step)