Practical guides
Ask a useful question to AI
A useful request identifies a result, the context that changes it and a way to check it. More instructions do not automatically mean a better answer; add only what helps the task.
Method
1. Define the result
Choose an action such as explain, compare or revise. State who will use the result and what they should be able to do afterwards. Separate a mandatory constraint from a preference.
2. Provide relevant context
Include the starting material, reader’s level and practical limits. Remove confidential details. If information is missing, ask the assistant to identify the gap rather than fill it with an invented fact.
3. Review one aspect at a time
Inspect the first answer against the objective. Keep useful parts and request a targeted change. A fluent answer still needs factual checking; a detailed prompt cannot guarantee correctness.
Example to adapt
Using these anonymised notes, prepare a 20-minute revision session for a beginner. Give four steps and three recall questions. Separate content in the notes from additional explanations and identify missing information.
Checklist
Checks stay on this page; ticking a box does not automatically validate the result.
Deliverable
A request containing the goal, context, constraints, format and acceptance criteria.
Limitations
Avoid adding personal details simply to make a prompt longer. Split unrelated objectives into separate requests.
Sources for further reading and checking
These independent resources help you find a reference. Check the language and scope: a French resource may not describe rules in another country. These are not sources automatically consulted for generated answers.
- NIST — Generative AI Profile (AI 600-1)
A generative AI risk profile for identifying and managing risks.
Language: en · Scope: US
Link check: 2026-10-06 · These links are reading resources. The answer service does not consult them automatically. A checked link does not guarantee every document is current.