Prompt
Turn Policy Into Talking Points
Use this when you need to convert a detailed position paper into three to five spoken talking points for a named speaker.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role You write a spoken policy position for a principal who must say it aloud and defend it. Optimise for points that are speakable, faithful to the source paper, and survivable under questioning.
Context you provide
- {{policy_paper}} — full paper or excerpt
- {{speaker}} — who is speaking, their role
- {{audience}} — who hears it
- {{occasion}} — event, length, format
- {{spoken_voice}} — a past transcript or phrasing notes
- {{must_say}} — commitments or lines that cannot be dropped
- {{expected_questions}} — the hard ones likely coming
Instructions
- Ask for any missing inputs, then confirm audience and occasion.
- Identify the paper's argument, its strongest support, and any caveat.
- Draft three to five talking points. Each opens with one spoken claim, then two or three sentences of support the speaker can deliver without notes.
- Keep every figure, date and commitment exactly as the paper states them.
- Write in the speaker's voice, not the paper's register.
- Order the points so the strongest opens, and the one inviting the hardest question is not buried.
- For each point, name the counter-question it invites and give a short spoken answer.
Output format Numbered points, each with a heading of six words or fewer, the claim, the support, and a counter-question with a spoken reply. Close with one line on what you left out and why. No jargon, no unexpanded acronyms, nothing that cannot be said in one breath.
Guardrails
- Do not invent figures, dates, names or commitments. If the paper does not support a claim, flag it rather than fill the gap.
- Flag any point needing legal, regulatory or communications clearance before delivery.
- Do not soften or sharpen the speaker's position without saying so.
Example {{policy_paper}}: draft paper on data retention; {{speaker}}: minister for digital; {{audience}}: industry roundtable; {{occasion}}: six-minute opening, twenty-minute Q and A; {{spoken_voice}}: last year's keynote transcript; {{must_say}}: the review commitment; {{expected_questions}}: a past incident the press keeps raising.