How the prompt works
The system prompt sets the LLM up as a communication consultant and requires six improvements in every rewrite: preserved detail, improved clarity, matched tone, better structure, added context, and clear action items. Without that structure, models either over-simplify (losing key technical detail) or produce generic rewrites that don’t fit the audience. The user prompt takes five variables — original message, target audience, desired tone, context, and primary goal — so every rewrite is shaped by who needs to read it and what it needs to achieve.System prompt
Sets the LLM’s role and defines the six improvements it must apply in every rewrite.User prompt
Five variables that shape the rewrite for the right audience, tone, and outcome.Sample output: API rate limiting issue
Sample output: API rate limiting issue
Quick update: API issue impacting Q2 launch timelineHi team,I wanted to loop everyone in on a technical issue affecting our Q2 release timeline and get us aligned on next steps.What’s happening:
Our API is hitting rate limits during peak usage times, causing connection errors for our mobile app. Think of it like a highway during rush hour — we’re getting more traffic than our current lanes can handle.The numbers:
- Our system can handle 1,000 requests per minute
- We’re seeing 1,200–1,500 requests during busy periods
- This is preventing our mobile team from completing their final testing
- Quick fix: Increase our rate limits to handle current traffic
- Better fix: Implement smart retry logic so the app handles busy periods gracefully
- Bonus: Optimise database connections for better performance
- Engineering will implement option 1 by Wednesday for immediate relief
- Option 2 is planned for next sprint as a permanent solution
- Mobile team can resume testing by Thursday