A polite public service announcement
If you wouldn't read it, don't expect me to. Generating text is cheap. Reading it is not. Own your output before you hit send.
Two messages. Same intent. One respects the reader.
Research summary01
Don't
Here's the AI research brief on pricing expansion.
[5 pages of market sizing, competitor tables, and SWOT analysis]
Do
For the June launch, I wouldn't add more pricing tiers yet.
The research points to confusion in the current plan, not demand for extra options, so I'd simplify the pricing page first.
Debugging help02
Don't
AI wrote this SQL query. Why doesn't it work?
[120 lines of schema, prompt, and generated query]
Do
This SQL query duplicates rows after joining orders to line_items.
I need one row per order with total quantity. What join or grouping am I missing?
Relevant query attached.
Executive answer03
Don't
Hey! Great question. Let me unpack this comprehensively.
[4 paragraphs of AI-generated text about stakeholder alignment, strategic synergies, cross-functional paradigms, and transformative outcomes]
Do
Yes, but only if we lock the budget by Friday.
Two risks I'd flag: vendor lock-in, and the migration window clashing with Q3 launch.
I drafted with AI then cut it to this. Happy to expand.
01
When you forward something you haven't read, you're telling me my time is worth less than yours.
02
A six-paragraph response to a yes/no question isn't thoroughness. It's noise.
03
Whatever tool you used to generate it, once it leaves your fingers, it's from you.
04
If you can't summarize it in your own words, you're not ready to send it.