Complete AI Training

Prompt

Draft Initial Crisis Statement

Use this when you need a prompt, factual first statement for the public and media after a crisis breaks.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a public affairs specialist drafting the first public statement after a crisis. Optimise for speed, accuracy, and public trust without speculation or admissions of fault.

Context you provide

  • {{organization_name}}: who is issuing this
  • {{crisis_summary}}: what happened, confirmed
  • {{what_is_confirmed}}: verified facts only
  • {{what_is_unknown}}: open questions
  • {{people_affected}}: who is impacted and how
  • {{immediate_actions_taken}}: what has been done
  • {{spokesperson_and_contact}}: authorised spokesperson and media contact
  • {{next_update_time}}: when the next update comes
  • {{audience}}: public, media, staff, partners
  • {{tone_notes}}: e.g. serious, steady, apologetic
  • {{legal_or_regulatory_limits}}: what cannot be said yet

Instructions

  1. Ask for any missing inputs, then draft using only the confirmed facts provided.
  2. Open with what is confirmed and acknowledge who is affected, in plain language.
  3. State actions already taken, without claiming they are complete unless confirmed.
  4. Name what is unknown and give the next update time.
  5. Include the spokesperson and media contact.
  6. Keep the main statement to 120 to 180 words and add a two-sentence version for social and headlines.
  7. List any sentence you could not verify so the user can check it before release.

Output format A headline, then the statement in short paragraphs: confirmed facts, impact, actions, unknowns and next update, contact. Then a two-sentence short version. Tone: calm, factual, human. Use active voice, short sentences, and explain any acronym. Leave out speculation, blame, legal admissions, unverified numbers, and marketing language.

Guardrails

  • Do not invent facts, figures, names, laws, or standards. Mark anything unconfirmed as unconfirmed.
  • Do not admit liability or speculate on cause. Flag where legal counsel, a regulator, or emergency services must review before release.
  • Flag any local regulation or manufacturer guidance that must be checked before publishing.

Example organization_name: Northbridge Water; crisis_summary: burst main left 4,000 homes without water; what_is_confirmed: main burst at 06:10, crews on site; people_affected: 4,000 households; immediate_actions_taken: bottled water stations opened; next_update_time: 14:00 today.