If AI keeps giving you bland, generic answers, the problem usually isn’t the tool. It’s the question. One-line prompts get one-size-fits-all replies. The fix is to stop asking AI things and start briefing it, the way you’d brief a capable new colleague who knows nothing about your situation.
Why one-line prompts disappoint
An AI model fills every gap you leave with the most typical, average answer it can produce. Ask it to “write an email about a delay” and it has to guess who the email is to, what was delayed, how serious it is, what tone to take and how long it should be. Its guesses are reasonable, which is exactly why the result sounds like everyone else’s.
Every detail you add removes a guess. That’s the whole skill.
The five-part brief
A good brief answers five questions. You don’t need all five every time, but the more that matter to the task, the better the result.
- Role: who should the AI act as? “You are an experienced account manager at a small IT support company.”
- Task: what exactly do you want produced? “Write an email to a client explaining that their server migration will be delayed by one week.”
- Context: what does it need to know? “The delay is caused by a supplier shipping the wrong hardware. The client is generally happy with us but has a busy month-end.”
- Format: what should the output look like? “A short email: subject line, three short paragraphs, then a clear next step.”
- Constraints: what should it avoid or stick to? “Under 150 words. Apologise once, not repeatedly. No jargon. Don’t blame the supplier by name.”
Before and after
Before: “Write an email to a client about a delay.”
After: all five parts above, in one message. The result is an email you could send after a quick read-through, rather than a template you’d have to rewrite.
Give it examples of what good looks like
The fastest way to get your tone right is to show it. Paste in a previous email you were happy with (anonymised; see our Data Red-List guide) and add: “Match the tone and length of this example.” One good example often does more than a paragraph of description.
Refine, don’t restart
The first answer is a draft, not a verdict. Instead of starting again with a new prompt, tell the AI what to change:
- “Make it warmer and cut the second paragraph.”
- “Too formal. This client is on first-name terms with us.”
- “Give me three alternative subject lines.”
- “What’s missing that the client is likely to ask about?”
Two or three short follow-ups usually get you from “nearly” to “ready”.
Always check before you send
AI writes fluently, and fluent text is easy to trust. Before anything leaves your hands, check four things: facts (dates, figures, commitments), names (spelt right, the right person), promises (did it commit you to something you can’t deliver?) and tone (would you say this out loud to this person?).
Build your own prompt library
When a brief works well, save it. Replace the specific details with placeholders like [CLIENT], [ISSUE] and [DEADLINE], and keep it somewhere you can find it again. In a few weeks you’ll have a personal library of briefs for the tasks you do most, and the time savings start to compound.
