Vendor-neutral · Beginner · 5 lessons · about 28 minutes · updated 2026-09-12
How to Write a Good AI Prompt: Talk to AI Without Jargon
One three-part habit, five short lessons, no vocabulary to memorise. Get a usable answer on the first try more often, and know exactly what to say when you do not.
For anyone who has used ChatGPT, Claude or Gemini and thought the answers were fine but not quite what they wanted. No technical background, no prompt-engineering vocabulary. If you have never used AI at all, take AI From Zero first; it is 30 minutes.
Course outline
- 1Task, Context, Output: the only three parts a prompt hasEvery good prompt says what to do, what to work with, and what shape to answer in. See a weak email prompt and a better one side by side, then run the better one.5 min · 3 objectives
- 2See it, copy it, change itThree fill-in prompts with less and less written for you. Start by changing one word, finish by writing all three parts yourself. By the end the habit is yours.6 min · 3 objectives
- 3Giving AI what it needs to knowThe model cannot see your screen, your inbox or your situation. Learn what to paste, what to state, and what to forbid, then run a prompt that shows how much context changes the answer.6 min · 3 objectives
- 4Asking for the shape you wantBullets, a table, a word limit, a tone. Models default to medium-length prose unless you say otherwise. Learn the handful of output words that work everywhere, and run a prompt with a real word limit.6 min · 3 objectives
- 5Fixing a weak answerThe first answer is a draft. Learn the four follow-ups that fix most of them, ask the model what it assumed, and practise telling a trustworthy answer from a confident one.5 min · 3 objectives
Frequently asked
Do I need to learn prompt engineering to use AI well?
No. For everyday use, one habit covers it: say what to do, what to work with, and what shape to answer in. This course teaches that habit and the follow-ups for when the first answer is not right.
Does this course work for ChatGPT, Claude and Gemini?
Yes. Task, Context, Output is the same on every model. The examples run in a Try it box on the page; you can paste any of them into any chatbot and get a similar result.
How long does it take?
Under 30 minutes for the five lessons, including running the examples. Each lesson is five or six minutes and works on its own.
Will better prompts save me money?
Usually. Most people who upgrade to fix weak answers were missing context or a format instruction, which are free. On the API, asking for short, fixed-shape answers is also the biggest single cost saver.