Library › Marketing AI › Marketing and content
Compare features with a rival and mark what you can't confirm
Practice: Catching mistakes and made-up facts. About 5 minutes.
About this exercise
You're building a feature comparison of Plotline (your project app for architects) against rival Draftboard. Write exactly 4 numbered lines in the shape Feature: Plotline yes/no | Draftboard yes/no/unknown, using Draftboard's pricing page as the only source for them. Anything not on their page is "unknown", and reply with only the plain-text list.
Common mistakes when you ask AI for this
- Filling gaps with rumors. The Reddit comment and the rep's hunch look like answers. Try: Add: "Only their pricing page counts. Otherwise say unknown."
- Overstating your own product. The AI marks offline mode yes because it's planned. Try: Add: "Plotline offline mode is no."
- No line shape. The AI builds a table with checkmarks and bold headers. Try: Show it: "Feature: Plotline yes | Draftboard unknown."
- Extra commentary. The AI adds a summary of who wins. Try: Add: "Exactly 4 numbered lines, nothing else."
What a strong prompt includes
- Limits sources
- Uses unknown
- Sets the line shape
- Says how many
- Asks for only the list
Related exercises
- Alert sales to a competitor's price change, facts only (Research AI · Product managers)
- Announce a referral bonus with the steps to claim it (Marketing AI · HR and recruiting)
- Answer an angry 'I unsubscribed' email honestly (Marketing AI · Marketing and content)
- Answer your CMO's campaign question in one Slack message (Marketing AI · Marketing and content)
- Approve or return expense claims against the policy (Productivity AI · Finance and accounting)