Beginner Level

What it is: Large Language Models are neural networks trained on billions of texts to understand and generate language.

Why: Understanding how the model "thinks" lets you create effective prompts.

Concepts: Token prediction, attention, temperature, hallucinations.

What it is: Tokens are pieces of text that the model processes. The context window is the total limit.

Why: Knowing how to count tokens helps you optimize prompts and avoid truncation.

Concepts: ~4 characters = 1 token, context windows from 4K to 200K+, cost per token.

What it is: Ideal structure: context/role, instructions, input data, output format.

Why: Well-structured prompts produce more consistent responses.

Concepts: System vs. user, delimiters, examples, constraints.

What it is: Be specific and direct, eliminating ambiguity.

Why: The clearer your request, the better the response.

Concepts: Avoid vague pronouns, specify the format, define the length.

What it is: Ask the model to perform a task without providing examples—just clear instructions.

Why: The simplest and fastest technique. Works for familiar tasks.

Concepts: Direct instructions on when to use and when to avoid.

What it is: Provide 2-5 examples before asking for the response. The model learns the pattern.

Why: Dramatically improves quality for specific tasks.

Concepts: Representative examples, consistency, ideal quantity (3–5).

What it is: Ask the model to "think aloud," showing each step of its reasoning.

Why: Dramatically improves responses in logic, math, and decision-making.

Concepts: "Think step by step," reasoning validation.

What it is: Give the model a role: "You are an expert in...".

Why: Automatically adjusts tone, vocabulary, and technical depth.

Concepts: Defining expertise, tone of voice, and role limitations.

What it is: Techniques for making text sound human-written.

Why: Increases authenticity and engagement. Avoids an “AI tone.”

Concepts: Contractions, personal opinions, structural variation.

What it is: Use AI to create drafts, check grammar, and adjust tone.

Why: Speeds up content production by 5–10x while maintaining quality.

Concepts: Tone of voice, target audience, iterative review.

What it is: Prompts for emails: follow-ups, negotiations, requests.

Why: Email is 40% of the work. The right prompts save hours.

Concepts: BLUF structure, formal/informal tone, call to action.

What it is: Summarize documents, articles, and meetings. Extract key points.

Why: Processing information is the bottleneck of the 21st century.

Concepts: Summary levels (executive, detailed), insight extraction.

What it is: Use AI as a brainstorming partner: generate ideas and explore angles.

Why: AI doesn't have creative block. An infinite idea generator.

Concepts: Divergence/convergence, SCAMPER, "what if...".

What it is: Instructions for AI to format with headings, bullets, tables, and code.

Why: Well-formatted answers are easier to read and use.

Concepts: Headers (#), lists (-), tables (|), code (```).

What it is: Ask for responses in a table to compare options.

Why: Tables make decisions easier. Side-by-side visualization is powerful.

Concepts: Relevant columns, sorting, CSV/Markdown format.

What it is: Use AI to create templates: emails, reports, proposals.

Why: Templates save time. Create once, use infinitely.

Concepts: [FIELD] placeholders, variations, personalization.

What it is: List of the most frequent errors: vague prompts, no context, too long.

Why: Learning from other people’s mistakes accelerates your learning curve.

Concepts: Ambiguity, lack of examples, failure to iterate.

What it is: Know the limits: hallucinations, cutoff date, mathematical reasoning.

Why: Realistic expectations prevent frustration.

Concepts: Hallucinations, bias, fact-checking.

What it is: A compilation of tips from experienced prompt engineers.

Why: Years of experience in immediately actionable tips.

Concepts: "Take a deep breath," ideal temperature, when to restart.