Prompting & Grounding
Prompting with context
Write prompts that give AI the audience, purpose, constraints, tone, and output format it needs.
5 min readPrompting
Workplace example
Better policy prompt
Instead of "Explain this policy", ask: "Explain this policy for a non-technical manager in 120 words. Include three risks, one recommended action, and avoid legal advice."
What this means
- •A useful prompt explains the job the output needs to do, not just the topic.
- •Good workplace prompts usually include audience, purpose, context, key points, tone, constraints, and desired format.
- •The prompt should also say what the AI should not do, especially where facts, sources, or sensitive information matter.
Why it matters
- •Vague prompts produce generic answers that may not fit the business context.
- •Clear prompts reduce rework and make outputs easier to review.
- •Prompting is not magic wording. It is structured communication.
Common mistakes
- •Asking "write something about..." without audience or purpose.
- •Asking for polish before checking accuracy.
- •Adding context only after the output is already weak.
What good judgement looks like
- •State the audience and decision the output supports.
- •Give enough context for the AI to avoid generic advice.
- •Specify the output structure so review is easier.
Try this at work
- •Rewrite one vague prompt using audience, purpose, context, constraints, tone, and format.
- •Compare the output from the vague prompt and the improved prompt.
- •Note what still needed human editing.
How this helps your reassessment
- •You can improve a vague prompt.
- •You know which details help AI produce a more useful workplace draft.
- •You do not confuse formal wording with good prompting.