What is a prompt, and how do you ask a good question?
A prompt is the instruction or input we send to a model.
What is it?
A prompt is the instruction or input you send to a model: a question, a text, an image or a task description. A clear prompt says what you want, for whom, in what form and within what limits. A good prompt improves the answer; it does not make the model infallible.
An everyday example
Instead of “Write about the climate”: “In 120 words, explain the difference between weather and climate for a primary school pupil.”
Why does it matter?
A clear instruction saves time and avoids misunderstandings. What you leave out of a prompt matters too, such as sensitive personal data.
A common misconception
That there is a magic formula that guarantees a correct answer. Even the best prompt does not replace checking the facts.
In more detail
What should a good prompt include?
The goal, the circumstances, the data, the format of the answer and any limits. For example: “Summarise in five short sentences and flag anything unclear.”
Does a longer prompt guarantee success?
No. A clear, checkable question is often better than a very long, messy one.
Can I ask for a source?
Yes. But the model may invent a link. You have to open the source and check it.
What if the first result is poor?
Refine the goal, add an example, break the task into steps, or check it against another independent source.
Remember: A good prompt makes the answer more useful; it does not turn the model into an infallible source.
Reliable sources
Last reviewed 1 November 2026. An explainer, not professional, legal or medical advice. For rules and figures that change, check the official version.