10 AI prompts for tweets that earn replies

Generate 3 tweet variants asking an open-ended question about {topic}. Constraints: question must have multiple valid answers, be easy to answer in 1-2 sentences, connect to something the audience cares about. Avoid yes/no questions and 'what do you think?' Each tweet max 220 chars. Voice: curious, genuine. Format: 1-2 sentences of context + question.

Generate a tweet stating an opinion about {topic} that thoughtful people would disagree about. Constraints: the opinion should be defensible with evidence, not contrarian-for-contrarian-sake. Don't include a question at the end — let the disagreement provoke itself. Max 240 chars. Voice: confident but not aggressive. The opinion should split the audience roughly 60/40 on agree/disagree, not 95/5.

Generate a tweet inviting reciprocal sharing. Topic: {topic}. Format: 'I'll share mine first: [your contribution]. Tell me yours.' or 'What's your X? I'll go first: [your X].' Max 240 chars. The 'going first' element is critical — readers won't share unless you commit publicly first. Voice: genuine, curious.

Generate a tweet asking readers about their own experience with {topic}. Format: a short personal observation from your own experience, then the question. Max 230 chars. The observation grounds the question; the question invites participation. Avoid generic 'what's your experience?' — be specific about what aspect of {topic} you're asking about.

Generate a tweet stating a hot take about {topic}. Constraints: explicitly invite disagreement at the end ('What am I missing?' or 'Tell me where I'm wrong'). The take should be defensible but provocative. Max 240 chars. Voice: confident, open. The disagreement invitation prevents the take from feeling closed-off.

Generate a tweet posing a quick self-test about {topic}. Format: 'Quick test: can you do [X], [Y], [Z]? If yes, you're in the top {N}% of {target audience} for {skill}.' Max 240 chars. The test should be specific (3-4 concrete checks), achievable for the audience, and the percentile claim should be defensible.

Generate a tweet using a confession-style opener about {topic}. Format: 'Confession: I {do/did/believe} {something most people in {field} don't admit}.' Then a brief defense. Max 230 chars. The confession should be specific and lightly self-deprecating but not self-destructive. Voice: honest, not performative.

Generate a tweet sharing a frustration about {topic} that the audience likely shares. Format: state the specific frustration in 1-2 sentences, then invite agreement or counterpoint. Max 240 chars. Voice: matter-of-fact, not whining. The frustration should be specific (not 'people are dumb') and one the audience will identify with.

Generate a tweet revealing a result from a survey/research you did. Context: surveyed {N} {audience}, asked about {question}, the surprising finding was {finding}. Format: 1) Sample + question. 2) Surprising finding. 3) Open question inviting reader interpretation. Max 240 chars. Voice: researcher reporting back.

Generate a tweet posing a two-option question about {topic}. Format: 'Quick: would you rather {option A} or {option B}? Wrong answers below.' Max 220 chars. Constraints: both options must be defensible. Voice: playful but substantive. The 'wrong answers' phrasing invites contrarian replies; the 'quick' framing makes it easy to participate.