Prompt Engineering 101: Get Better Results from AI

Prompt engineering is the skill of communicating with AI models in a way that produces useful, accurate, and professional results.

Most people use AI the same way they use a search engine — they type a short query and hope for the best. That approach works for finding restaurants. It doesn't work for getting AI to write a client proposal, build a spreadsheet formula, or analyze a dataset.

The difference between a mediocre AI output and a great one is almost always the prompt. AI models are extremely capable, but they need clear, specific, and well-structured instructions to perform at their best.

Prompt engineering is not a technical skill. You don't need to know how AI models work under the hood. You just need to learn how to communicate clearly — and apply a few proven techniques.

This course teaches you the fundamentals. By the end, you'll get consistently better results from any AI tool you use.

Every strong prompt has the same four ingredients. You don't need all four every time, but knowing them makes you better at writing any prompt.

1. ROLE Tell the AI who to be.

"You are an experienced copywriter who specialises in B2B SaaS." "You are a senior financial analyst." "You are a hiring manager reviewing a resume."

Giving the AI a role primes it to draw on the right knowledge and tone. Without a role, you get a generic, average answer.

2. TASK Tell the AI exactly what you want it to do.

Be specific. "Write a blog post" is weak. "Write a 600-word blog post introduction about remote work for small business owners, written in a conversational tone" is strong.

3. CONTEXT Give the AI the information it needs.

If you're writing a job post, include: the role, the company, the skills required, and the tone. If you're writing a reply to a client complaint, paste the complaint in. Context is the difference between generic output and output that's actually useful.

4. FORMAT Tell the AI how you want the output structured.

"Give me a numbered list." "Format this as a table with columns for: Task, Owner, Deadline." "Write this as a professional email. Keep it under 150 words." "Give me 5 options."

When you include all four ingredients, AI produces better output faster — and you spend less time editing.

These techniques work whether you're using ChatGPT, Claude, Gemini, or any other large language model.

CHAIN OF THOUGHT Ask the AI to think step by step before giving you an answer.

"Before answering, think through this step by step."

This dramatically improves accuracy for anything that involves reasoning, analysis, or multi-step problems.

FEW-SHOT EXAMPLES Show the AI an example of what you want before asking it to produce the real thing.

"Here is an example of the type of headline I want: [example]. Now write 10 headlines for [product] in the same style."

Examples are more powerful than descriptions. Showing beats telling.

ITERATIVE REFINEMENT Don't expect perfection on the first try. Prompt in layers.

First prompt: Get the rough draft. Second prompt: "Make the tone more professional." Third prompt: "Cut this down to 200 words." Fourth prompt: "Add a specific example in paragraph 2."

Working iteratively is faster than trying to write one perfect prompt upfront.

NEGATIVE INSTRUCTIONS Tell the AI what not to do.

"Do not use bullet points." "Do not use the word 'leverage'." "Do not include any preamble — start directly with the content."

AI models have default tendencies. Negative instructions override them.

PERSONA PROMPTS FOR FEEDBACK Ask AI to give you critical feedback by putting it in a specific role.

"You are a demanding client who is never satisfied with the first draft. Review this proposal and tell me exactly what you would push back on."

This surfaces weaknesses before you send something to a real client.

Here are prompt frameworks for the tasks remote professionals do most often.

WRITING "You are an experienced [copywriter / technical writer / content strategist]. Write a [type of content] for [audience] about [topic]. The tone should be [adjective]. Keep it under [word count]. Do not use filler phrases or corporate jargon."

RESEARCH & ANALYSIS "You are a senior research analyst. Summarise the key trends in [industry/topic] in 2025. Structure your response with: (1) 3–5 main trends, (2) one real-world example for each, (3) what this means for [specific audience]. Be specific — avoid generalities."

EMAIL & COMMUNICATION "Draft a [professional / friendly / firm] email to [recipient] explaining [situation]. The goal is to [outcome]. Keep it under [X] words. Do not start with 'I hope this email finds you well'."

DATA & SPREADSHEETS "I have a spreadsheet with columns: [list columns]. Write an Excel formula that [describes what you need]. Explain what the formula does in plain English."

CODE "You are a senior [language] developer. Write a function that [describes task]. Include comments. Handle edge cases for [specific scenarios]. Use [library/framework] conventions."

SOCIAL MEDIA "Write 5 LinkedIn posts about [topic] for [audience]. Each post should be under 200 words, start with a hook sentence, and end with a question or call to action. Use a conversational, direct tone. No hashtag overload."

Most people who are frustrated with AI output are making one of these mistakes.

MISTAKE 1: Being too vague "Write me a blog post about marketing" produces garbage. "Write a 700-word blog post for e-commerce store owners on how to reduce cart abandonment using email retargeting. Beginner audience. Include 3 specific tactics with examples." produces something useful.

Fix: Add specificity to every element — who, what, how long, what tone, what format.

MISTAKE 2: Not giving context AI doesn't know your client, your brand, your audience, or your situation. Without context, it invents generic defaults.

Fix: Always paste in relevant context. Client brief, existing copy, the email you're replying to, your company description. More context = better output.

MISTAKE 3: Accepting the first output The first draft is a starting point, not the final product. Most people give up after one try.

Fix: Always refine. "Make this shorter." "Make this more direct." "Add a specific example." Iteration is how you get to great.

MISTAKE 4: Asking one question that should be ten "Write me a full business plan" is too big for one prompt. Break it into parts.

Fix: Break complex tasks into steps. Prompt for each section separately. Then combine.

MISTAKE 5: Not checking outputs AI confidently states wrong information. Always verify facts, statistics, and anything that could embarrass you with a client.

Fix: Treat AI output as a draft, not gospel. You are still responsible for accuracy.