My GenAI Prompt Engineering Cheatsheet

I created the following cheat sheet for my students and thought I'd share broadly. Following are the different prompt engineering frameworks (columns) and use case the are good for (rows).

CLEAR Framework

Structured approach for concise and effective prompts.

• Context: Provide necessary background information.
• Limitations: Define scope and constraints.
• Examples: Provide sample outputs or expected format.
• Assertiveness: Use direct and specific language.
• Refinement: Iterate based on response quality.

Example Prompt: "You are an AI trained in investment strategies. Provide a summary of risk-adjusted returns for private equity investments. Use bullet points. Limit to 5 key insights."

• Best for: Business Strategy & Decision-Making Optimization & Refinement Generating Executive Summaries
• OK for: Research & Analysis Breaking Down Complex Topics

CRISP Framework

Best for getting refined, well-structured responses.

• Context: Describe the problem or scenario.
• Role: Assign a specific persona or expertise.
• Instructions: Define what you expect.
• Scope: Narrow down the focus (length, format, depth).
• Parameters: Set boundaries (e.g., "keep it high-level" or "use layman’s terms").

Example Prompt: "Act as a senior financial analyst. Explain how to evaluate AI-driven investment models. Keep it high-level and limit to 200 words."

• Best for: Optimization & Refinement Business Strategy & Decision-Making
• OK for: AI & Technical Explanations Generating Executive Summaries

PROMPT Framework (by OpenAI Community)

A step-by-step method for refining prompts.

• Persona: Assign an AI role (e.g., CTO, economist, historian).
• Request: Clearly state what you need.
• Outcome: Specify expected output type (table, summary, analysis).
• Metrics: Define success criteria (accuracy, conciseness, etc.).
• Polish: Iterate for better quality.
• Tone: Set the style (formal, casual, technical).

Example Prompt: "You are an AI expert writing for busy executives. Summarize the impact of AI on private equity. Keep it under 150 words, with a focus on efficiency gains."

• Best for: Research & Analysis Business Strategy & Decision-Making Generating Multiple Options
• OK for: Optimization & Refinement

Chain of Thought (CoT) Prompting

Encourages logical step-by-step reasoning.

• Ask for intermediate reasoning before providing the final answer.
• Best for complex problem-solving, coding, and logical deduction.

Example Prompt: "Explain how AI-driven credit risk modeling works. Break it down step by step, starting with data collection, then feature engineering, then model selection, and finally risk assessment."

• Best for: AI & Technical Explanations Breaking Down Complex Topics
• OK for: Research & Analysis Business Strategy & Decision-Making

Few-shot & Zero-shot Prompting

• Zero-shot: Direct query without examples.
• Few-shot: Provide examples to guide the response style.

Zero-shot Example: "Summarize the latest trends in AI governance."

Few-shot Example:

Summarize AI governance trends like the following examples:

Explain EU’s AI Act in 3 bullet points.

Summarize US AI regulations in 2 sentences.

Now, do the same for China's AI policies.

• Best for: Research & Analysis Generating Multiple Options
• OK for: Business Strategy & Decision-Making Generating Executive Summaries

RACE Framework (for decision-making prompts)

Useful for executive and analytical insights.

• Reasoning: Ask the AI to explain the logic behind its response.
• Alternatives: Request multiple perspectives or options.
• Considerations: Include risk factors or dependencies.
• Evaluation: Ask for pros and cons.

Example Prompt: Compare agentic AI models with traditional AI models. Provide reasoning, key differences, risks, and a decision matrix for investment.

• Best for: Business Strategy & Decision-Making Generating Multiple Options
• OK for: AI & Technical Explanations Research & Analysis

Role-based Prompting

Assigning a role makes responses more relevant.

Example Prompt: As a CTO, how should I assess SLA levels for AI-driven infrastructure? Provide a checklist.

• Best for: AI & Technical Explanations
• OK for: Business Strategy & Decision-Making Engaging Conversations

Self-Reflective Prompting

AI evaluates and improves its response.

Example Prompt: Answer the following question: 'What are the key AI trends in financial services?' Now, critique your own response for completeness and bias, and refine it.

• Best for: Optimization & Refinement Self-Improving AI Responses
• OK for: Business Strategy & Decision-Making AI & Technical Explanations

Contrastive Prompting

Compare two different perspectives.

Example Prompt: "Compare the benefits and risks of agentic AI in investment strategies. Present both sides objectively."

• Best for: Comparing Different Perspectives
• OK for: Research & Analysis Business Strategy & Decision-Making

Interactive Prompting

Use back-and-forth questioning to refine answers dynamically.

Example Prompt: "What are the best strategies for implementing AI in investor relations? Ask me follow-up questions before giving a response."

• Best for: Engaging Conversations
• OK for: Self-Improving AI Responses Comparing Different Perspectives