Design a safeguarding training scenario
Create a realistic, non-triggering safeguarding scenario for staff training.
This one needs human approval before use — got it?
High-risk prompt: touches sensitive areas. Always have a human review the output before you use it. How to use AI responsibly →
Fill in the details
2 placeholdersType your details in and we'll drop them into the template. Empty slots stay as tokens so nothing goes missing.
Prompt template
Design a fictional safeguarding scenario for a training session for {{association_name}} staff working with {{communities_served}}. Requirements: - The scenario must be entirely fictional; do not use any specific real cases - It must not describe abuse in graphic terms - It should raise a realistic ambiguity — the "what would you do?" moment - Include the moment a staff member has to decide whether to escalate - Provide 4 discussion questions - Provide the recommended response (what the safeguarding lead would want to see) Do not include names, locations, or details resembling any real situation.
Replace each highlighted placeholder with real context. Keep names and safeguarding details out.
Human review checklist
Tick each before you send or publish the output.
When to use this prompt
Reach for this prompt when you need to create a realistic, non-triggering safeguarding scenario for staff training. It's designed for Safeguarding lead, Ops lead. Typically used within Safeguarding work.
Not suitable for real-time decisions about individuals, safeguarding cases, or public claims without human sign-off.
How to use it
- 1
Gather your inputs
You'll need: communities_served. - 2
Fill in the highlighted placeholders
Copy the template above and replace every highlighted placeholder with real, local context. - 3
Run it in your AI tool
Paste into your preferred AI assistant. Ask a follow-up if the output misses your association's tone. - 4
Work through the checklist
Tick every item before you send, publish or act on the output.
Why this prompt works
Explicit content constraints prevent the model from importing distressing real cases into a training context.
Was this useful?
Sign in to leave feedback and shape which prompts stay in the library.