Rewrite app messages so they help instead of blame

About this exercise

You're the PM for Pocketpal, a budgeting app, and the app's messages sound robotic. Rewrite the 4 messages as exactly 4 numbered lines, each under 14 words, each saying what to do next. Follow the voice guide in your task: never blame the user (no "you failed", "invalid" or "error"), no "oops" or "uh-oh", and no exclamation points. Plain text, only the list.

Common mistakes when you ask AI for this

• Saying "make these friendlier". The AI adds "Oops!" and exclamation points, which the team already rejected. Try: Add: "No oops, no uh-oh, no exclamation points."
• Keeping the blame words. The AI softens tone but keeps "invalid" and "failed". Try: Add: "Never use error, invalid, failed or rejected."
• Not saying what to do. The AI explains the problem without the next step. Try: Add: "Each message says what to do next."
• Getting a table with notes. The AI shows before/after with bold labels and rationale. Try: Add: "Exactly 4 numbered lines, plain text, no notes."

What a strong prompt includes

• Says how many
• Asks for the next step
• Bans blame words
• Rules out oops and exclamation points
• Asks for only the list

Practice this now ↑

Related exercises

• Announce a referral bonus with the steps to claim it (Marketing AI · HR and recruiting)
• Answer a recruiter's message briefly and confidently (Writing AI · Job seekers)
• Answer an angry 'I unsubscribed' email honestly (Marketing AI · Marketing and content)
• Answer an upset parent's grade question with the math first (Writing AI · Teachers)
• Answer your CMO's campaign question in one Slack message (Marketing AI · Marketing and content)