Necessary
Required for secure sessions, sign-in, CSRF protection, language selection and saving your cookie choice.
← Back to blog
Four parts of a clearer AI request
State the goal, context, limits and output format to make the response easier to review.
A longer message does not automatically produce a better AI answer. Irrelevant detail can be as unhelpful as missing context. A useful request explains the task and makes the result checkable. Even a well-written prompt cannot guarantee accuracy.
State the goal in one sentence
“Make this appointment reminder shorter and more polite” is clearer than “Improve this text.” It reduces how much the system must guess about your intention.
Add only necessary context
Audience, language and intended use may matter. A person's name, health history or identity number is unnecessary for many writing tasks. Use fictional examples where possible. Removing a name is not always enough to anonymise information; combined details can identify someone too.
Specify limits and format
For example: “Use three short sentences and add no price or outcome promises.” Ask for questions when information is missing rather than invented details. An expert profile should use only qualifications and services you have verified.
Compare the result with the task
Check added claims before admiring the style. Does the answer introduce a deadline, fee or guarantee absent from the source material? Request a focused correction if necessary, then review again. The most polished text is not always the most reliable. Final publication and decision-making should remain with a person.
Sources and further reading
NIST — Generative Artificial Intelligence Profile (AI 600-1)