AI Prompting Techniques Checklist

Core Fundamentals

0/8

Master the "Manager" approach: Create hyper-specific, detailed prompts that clearly define role, task, output, and constraints.

Assign clear roles using persona prompting to establish context, tone, and expected expertise.

Clearly outline the task and provide step-by-step plans for complex operations.

Implement escape hatches: Instruct the LLM to say "I don't know" rather than hallucinating.

Use debug info and thinking traces to understand the LLM's reasoning process.

Understand different model "personalities" and adapt prompting style accordingly.

Practice prompt distillation: Use larger models for optimization, then deploy on smaller models.

Create comprehensive prompts that treat the LLM like a new employee with detailed instructions.

Structure & Formatting

0/6

Use Markdown formatting (headers, bullet points) to structure prompts clearly.

Implement XML-like tags for structured outputs and machine-readable responses.

Define expected output format clearly with examples and templates.

Create consistent prompt templates for recurring tasks and workflows.

Use section headers and clear delineation between different prompt components.

Structure constraints and guidelines in easily parseable format.

Examples & Context

0/6

Provide high-quality few-shot examples with input-output pairs for complex tasks.

Use in-context learning to demonstrate desired style and format.

Include both positive and negative examples to clarify boundaries.

Create hard examples that challenge the LLM and improve performance edge cases.

Use examples to establish consistent voice, tone, and style.

Balance example quantity vs. quality - focus on representative, high-impact examples.

Advanced Techniques

0/7

Implement prompt folding: Design prompts that generate specialized sub-prompts dynamically.

Use dynamic generation for multi-stage workflows and context-aware prompting.

Master meta-prompting: Use LLMs to improve and critique your own prompts.

Create adaptive agentic systems with context-sensitive prompt generation.

Design classifier prompts that route to specialized downstream prompts.

Implement chain-of-thought reasoning for complex multi-step problems.

Use retrieval-augmented generation (RAG) patterns in prompt design.

Quality Control

0/4

Build comprehensive evaluation suites as your "crown jewels" for measuring prompt quality.

Create test cases that cover edge cases, failure modes, and performance metrics.

Implement systematic A/B testing for prompt variations and improvements.

Establish metrics for consistency, accuracy, and performance across different scenarios.

Optimization & Iteration

0/4

Use iterative refinement based on real-world performance and user feedback.

Monitor and analyze prompt performance in production environments.

Optimize for cost, latency, and quality trade-offs across different model sizes.

Document lessons learned and create prompt versioning for continuous improvement.

🤖 AI Prompting Techniques Checklist

Practice Prompt