AI prompt framework
Popular Prompt Frameworks and Techniques
1. ERA (Expectation, Role, Action)
• Expectation: Define what you want as the output.
• Role: Assign a role to the AI for context (e.g., teacher, marketer).
• Action: Specify what the AI should do.
Example: “As a history professor, summarize the causes of World War I in simple terms.”
2. APE (Action, Purpose, Expectation)
• Action: What task needs to be done?
• Purpose: Why is the task important?
• Expectation: Define the result you expect.
Example: “Create a social media post (Action) to promote sustainability (Purpose) with at least three actionable tips (Expectation).”
3. TAE (Task, Action, Goal)
• Task: Identify the specific task.
• Action: Describe the process or steps to complete it.
• Goal: State the desired outcome.
Example: “Write a blog article (Task) explaining renewable energy options (Action) to raise environmental awareness (Goal).”
4. STAR (Situation, Task, Action, Result)
• Situation: Set the context.
• Task: Specify the goal or problem.
• Action: Explain what steps were taken.
• Result: Provide the outcome.
Example: “A company had a drop in sales (S). The goal was to improve customer retention (T). They launched loyalty programs (A), resulting in a 20% sales increase (R).”
5. AIDA (Attention, Interest, Desire, Action)
• Attention: Grab interest with a bold statement or question.
• Interest: Provide engaging information.
• Desire: Showcase benefits or unique features.
• Action: Call for specific action.
Example: “Promote a fitness app by grabbing attention with health stats, sharing user success stories, and offering a 30-day free trial.”
6. TRACE (Task, Request, Action, Context, Example)
• Helps in clear client or team communication.
Example: “Create an email template (Task) requesting customer feedback (Action), including space for star ratings (Context).”
7. COAST (Context, Objective, Actions, Scenario, Task)
• Adds depth for complex project planning.
Example: “Set the objective of increasing app downloads in Q4, identify key actions (ads, campaigns), and outline scenarios for different budgets.”
8. PEAS (Performance measure, Environment, Actuators, Sensors)
• Often used for AI system design.
• Defines how AI interacts with its environment to achieve performance goals.
Techniques to Optimize Prompt Crafting
• Be Specific: Use detailed prompts to avoid ambiguity.
Example: “Write a 200-word summary of the benefits of AI in education.”
• Provide Context: Add background or situational details.
Example: “As a teacher, outline how AI tools can simplify grading tasks in schools.”
• Add Constraints: Include word limits, tone, or formatting preferences.
Example: “Write a 3-paragraph blog in a professional tone for healthcare executives.”
• Iterate and Refine: Experiment with rephrasing to get desired results.
• Use Role Assignment: Assigning roles like “expert,” “critic,” or “researcher” makes responses more tailored.
Summary of Key Points
• Frameworks:
• ERA, APE, TAE: Guide tasks with clear expectations and goals.
• STAR, AIDA: Focus on storytelling and engagement.
• TRACE, COAST: Useful for project or team-related prompts.
• PEAS: For AI and system design.
• Techniques:
• Be specific, set a role, and use constraints.
• Provide examples or references to guide AI.
• Break tasks into subtasks for better clarity.
Further Reading and Resources
1. Factspan: Detailed insights on frameworks like AIDA, STAR, and PEAS.
2. Unite.AI: OpenAI’s guide to enhancing prompts with specificity and context.
3. Shelly Palmer Blog: Simplified approaches for GPT usage in practical scenarios.
Shelly Palmer - Advanced Prompt Techniques
Explore these links for more examples and practical applications!