AI Freedom — Module 2 · Phase 1: Your AI Foundation
Prompt Engineering
That Actually Works
The difference between getting generic, useless output and getting something genuinely useful comes down almost entirely to how you write your prompt. This module teaches you exactly how.
What you will learn
By the end of this module you will be able to:
Write prompts that get useful output
Apply the four-part prompt framework to any task and produce output that requires minimal editing.
Give AI the context it needs
Understand why context is the single most important variable in prompt quality — and how to provide it efficiently.
Use iterative prompting
Refine output through conversation rather than trying to write the perfect prompt on the first attempt.
Build reusable prompt templates
Create and save prompt templates for the tasks you repeat most often, so you never start from scratch.
Lesson 1 — Why Most Prompts Produce Mediocre Output
The most common prompt pattern is: type what you want, press enter, feel vaguely disappointed with the result, try again with slightly different wording, feel slightly less disappointed, give up and edit it yourself. This pattern is not your fault. Nobody explained how these systems actually process instructions.
An AI model processes your prompt by attempting to generate the most statistically likely continuation of it, given everything it was trained on. When you write a short, uncontextualised prompt — “write me a cover letter” — the model has almost no information to work with. It does not know who you are, what role you are applying for, what your experience is, what tone the company uses, or how long the letter should be. So it generates the most generic, average version of a cover letter imaginable, because that is the most likely continuation of your prompt given its training data.
The fix is not to find a magic formula. It is to provide the information the model needs to do the job properly. Every good prompt is essentially a well-structured briefing document.
Lesson 2 — The Four-Part Prompt Framework
Good prompts consistently include four elements. You do not need all four for every task, but the more of them you include, the more useful the output becomes.
Part 1 — Role
Tell the AI who it is in this context
Not because AI has a personality, but because framing the role sets the tone, vocabulary, perspective and level of expertise you need. A prompt that starts “You are an experienced remote hiring manager” will produce very different output than one that starts “You are a career coach who has never hired anyone.”
Example: “You are an experienced remote recruitment specialist with 15 years of placing candidates in technology companies.”
Part 2 — Context
Give it the information it cannot know on its own
Who you are, what situation you are in, what you have already tried, what the audience is, what constraints exist. The more specific and relevant context you provide, the less the model has to guess — and guessing is where generic output comes from.
Example: “I am a marketing manager with 6 years of experience, currently applying for a fully remote content strategy role at a B2B SaaS company. The job description emphasises async communication and data-driven content.”
Part 3 — Task
State exactly what you want, in the format you want it
Be specific about the output format, length and structure. “Write a cover letter” is a task. “Write a 250-word cover letter in three paragraphs: one on why I want this specific role, one on my most relevant experience, one on why I would thrive working remotely” is a far better task.
Example: “Write a 250-word cover letter in three paragraphs. Do not start with ‘I am writing to apply for’. Use a direct, confident tone that signals remote work capability.”
Part 4 — Constraints
Tell it what to avoid
AI models default to certain patterns that are often unhelpful: starting with “Certainly!”, using corporate jargon, adding unnecessary caveats, using bullet points when paragraphs would work better. Stating constraints pre-empts these patterns.
Example: “Do not use phrases like ‘I am passionate about’ or ‘team player’. Do not add a closing paragraph asking them to read my CV — that is already implied.”
A complete example prompt
“You are an experienced remote recruitment specialist. I am a marketing manager with 6 years of B2B experience applying for a fully remote content strategy role at a SaaS company. The job description emphasises async communication and data-driven content.
Write a 250-word cover letter in three paragraphs: why I want this specific role, my most relevant experience, and why I thrive working remotely. Do not start with ‘I am writing to apply for’. Do not use the phrase ‘I am passionate about’. Use a direct, confident tone.”
Lesson 3 — Iterative Prompting: Refining Through Conversation
One of the biggest mistakes people make is treating each AI interaction as a single transaction. They write a prompt, get output, find it imperfect, and start a new prompt from scratch. This wastes the most powerful feature of conversational AI: memory within a session.
Within a single conversation, Claude remembers everything that came before. This means you can refine iteratively, treating the AI like a human collaborator: “That is good, but the second paragraph is too formal — make it more direct.” Or: “Shorten this by 30 words without losing the key points.” Or: “Give me three alternative versions of the opening sentence.”
Useful iterative prompts to save
• “That is close. The tone is slightly too formal — make it feel more like a confident professional, less like a corporate document.”
• “Keep the structure but rewrite the second paragraph to be more specific about results rather than responsibilities.”
• “Give me three alternative versions of the opening sentence.”
• “Cut 50 words without removing any key points.”
• “What is the weakest part of this? How would you improve it?”
• “Read this back as if you were the hiring manager. What impression does it create?”
Lesson 4 — Building Your Personal Prompt Library
The difference between someone who uses AI occasionally and someone who uses it systematically is a prompt library. A prompt library is simply a collection of prompts that work well for tasks you do repeatedly — saved somewhere you can find and reuse them.
Every time you write a prompt and the output is genuinely useful, save the prompt. Over time this library becomes one of your most valuable professional assets — it encodes your standards, your voice and your workflow in a form that consistently produces good results.
Organise your library by task type rather than tool. You will use different tools for different tasks, but the underlying task description transfers across tools. A good “summarise this document” prompt works in Claude, ChatGPT and Gemini with minimal adjustment.
Suggested prompt library categories
Writing & editing — drafts, rewrites, tone adjustments
Job search — CVs, cover letters, outreach messages
Research — company analysis, market research
Freelance — proposals, client emails, briefs
Daily work — meeting summaries, status updates
Content — LinkedIn posts, product descriptions
Module downloads
• The Four-Part Prompt Framework (reference card)
• 25 Prompt Templates for Remote Workers
• Iterative Prompting Cheat Sheet
• Prompt Library Template (Google Doc)
Your action before Module 3
Write three prompts using the four-part framework for tasks you actually do in your work or job search. Test each one in Claude. Save any that produce genuinely useful output to your prompt library. Notice which parts of the framework make the biggest difference to output quality.