AI Productivity
Ask AI This Before You Send a Work Message
Before you send a work message, paste it into an AI chat and ask: 'How could the reader misread this? Then give me one fix.' Copy-ready prompt and examples inside.
You write a quick message, hit send, and a minute later someone replies "Is everything okay?" The words were fine in your head. The reader heard something else.
Try the prompt from our video: before you send, ask an AI chatbot how your message could be misread, and suggest one fix.
The prompt
Paste your draft into a chatbot such as ChatGPT, Google Gemini, Claude or Microsoft Copilot, then add:
How could the reader misread this? Then give me one fix.
That's it. You're not asking AI to rewrite everything. You're asking it to read your words the way someone else might, then suggest one change.
An example
Your draft:
We need to talk about the launch.
What the AI might point out: it can read like bad news, and the reader may worry until you explain.
One fix:
We need to talk about the launch. Good news, just two dates to confirm.
If those details are true, the added context may reduce ambiguity. Do not add “good news” or “two dates” unless they accurately describe the situation.
Messages worth checking
- Short messages to a manager or client. Short messages can leave room for guessing.
- Anything that says no, pushes back, or changes a plan. Tone can matter here.
- Feedback on someone's work. A neutral note can read as harsh in text.
- Late-night or rushed messages. If you typed it fast, check it before it goes out.
A quick walkthrough
- Write your message the way you normally would.
- Paste it into your chatbot with the prompt above.
- Read how it says the message could land. Ask yourself if that's a fair reading.
- Take the fix, or tweak it so it still sounds like you.
- Check that the revised message preserves your intent and that every added detail is true before you decide to send.
Make it more useful (optional)
- Say who's reading: "The reader is my manager" or "The reader is a client I've never met." This can help tailor the suggestion.
- Ask for the worst reading: "What's the most negative way someone could read this?"
- Keep it short: "Keep your fix under 25 words."
- Keep your voice: "Don't make it more formal than my original."
Before you hit send
AI may lack relevant context about your history with the reader, your team’s inside jokes, or the chat, unless you provide it or it is available through the tool’s context or memory. Treat its answer as a second opinion, not a verdict. Don't paste anything confidential or anything your workplace doesn't allow in outside tools, and double-check any names, dates, or facts the fix adds.
The takeaway
Check before you send. Paste your message into any AI chat and type: "How could the reader misread this? Then give me one fix." A second look may help you spot ambiguity before sending.
Want to give AI more context before it answers? Read The One-Line Prompt That Makes AI Ask Before It Answers. For a tricky tone (refunds, chasing payment, saying no), see The AI Prompt for a Polite but Firm Email.
Send this to a friend who lives in team chat, and like & follow @ai_tools_day for more AI tips.