Free guide

Learn Prompt Engineering
in 5 minutes.

What a prompt is, why most of them fail, and the 5-part framework that fixes it — with live examples on your own prompts.

Start here

What is a prompt?

A prompt is the message you send to an AI. When you open ChatGPT and type something, that text is your prompt. The AI reads it, processes it, and responds. Everything about that response — the quality, the tone, the relevance, the format — is shaped entirely by what you wrote.

Most people treat it like a Google search: a few keywords, a quick question. That works for search engines because search has been trained to guess what you mean. AI hasn't been trained to guess what you need. It takes your prompt literally.

Same AI. Same question. See the difference.

Typical prompt

help me with my presentation

The AI will produce something generic. You'll rewrite most of it.

Structured prompt

I'm a product manager presenting a new feature to our engineering team on Friday. Write a 3-minute verbal opening that covers the business context, why this matters now, and what we're asking the team to build. Conversational tone — no slides, just talking points.

The AI has everything it needs. The output is usable, not just readable.

Why it matters

What is prompt engineering?

Prompt engineering is the practice of writing prompts intentionally — not just typing what comes to mind, but structuring your request in a way the AI can act on precisely.

It sounds technical. It isn't. You don't need to know anything about AI models, training data, or machine learning. You just need to give the AI what it needs: who it's talking to, what exactly you want done, what rules to follow, and how to format the answer.

The AI is literal.

It responds to what you wrote, not what you meant. Vague in, vague out. The more specific you are, the more useful the response.

Same AI, different results.

Two people asking the same AI the same question get completely different answers — because their prompts are different. The model is identical. The prompt isn't.

You don't need a course.

Prompt engineering is just structured thinking. Once you know the 5 elements, you'll do it naturally — in under a minute.

The good news: there are exactly 5 things that make a prompt go from vague to powerful. Master these and you'll consistently get results you can actually use — from any AI, for any task.

The framework

The 5 elements of a great prompt

Every strong prompt has some combination of these. Most weak prompts are missing most of them.

R

ROLE

Who should the AI be?

C

CONTEXT

What's the background?

T

TASK

What do you need done?

C

CONSTRAINTS

Any rules or limits?

F

FORMAT

How should the output look?

Now try it on your own prompt.

Paste any prompt below — we'll score it and explain exactly what's missing.

Your prompt

write me a cold email to a startup founder
Fix this prompt →

Live Analysis

type a prompt first

← Type a prompt to analyse it

The quick cheat sheet

ElementAsk yourselfExample phrase
ROLEWho should the AI be?"You are a senior [profession] with expertise in [topic]..."
CONTEXTWhat's my situation?"I'm a [role] working on [project]. The goal is [outcome]..."
TASKWhat exactly do I need?"Write / List / Compare / Summarize / Draft [specific deliverable]"
CONSTRAINTSWhat should it avoid?"Keep it under 150 words. Avoid jargon. Tone: casual and direct."
FORMATHow should it look?"Respond in a numbered list / markdown table / JSON / email format."
Ready to level up?

Now let PromptIt do this automatically.

You've learned the 5 elements. PromptIt applies all of them to every prompt — automatically, personalised to your profile, in seconds. Lives inside ChatGPT as a Chrome extension.