Resources

AI governance for real-world teams

1

Define what 'safe' means for your organisation

  • Agree categories of data that must never be shared with public tools.
  • Set expectations on prompts, retention, and where outputs can be used.
  • Choose deployment options based on risk: public, enterprise, VPC, or self-hosted.
2

Make guidance role-based

  • Teachers/admin staff need different guidance than developers or leadership.
  • Use examples tied to real workflows: drafting, summarising, classification, and search.
  • Provide a simple escalation path when staff are unsure.
3

Roll out in small, measurable steps

  • Start with low-risk use cases and build confidence and understanding.
  • Log decisions and assumptions so governance stays coherent over time.
  • Train staff on both capability and limitation, especially hallucinations and data leakage.