Home / Knowledge base / AI & governance
AI use-case brief: a one-page template
Most failed AI projects start with "let's add AI" instead of a specific problem. Fill in this brief before anything gets built.
| Section | Questions |
|---|---|
| Problem | What task is slow, expensive or error-prone today? Who does it? How often? |
| Users | Who will use the AI output: customers, staff or both? |
| Inputs | What data or documents will the AI use? Who owns them? Are they current? |
| Output | What exactly should the AI produce: an answer, a draft, a classification, an action? |
| Boundaries | What must it never do or say? When must it hand off to a person? |
| Success measure | Which number should improve, and by how much? |
| Baseline | What is that number today? |
| Risks | What happens if it's wrong? Who could be affected? |
| Human oversight | Who reviews outputs, and who can switch it off? |
| Evaluation | How will we test it before launch: which questions, what pass mark? |
| Owner | Who is accountable after launch? |
How to use it
- Keep it to one page.
- Get sign-off from the business owner and the team that will use it.
- Revisit it after the first month with real results.
A clear brief also feeds directly into AI governance documentation, such as the impact assessments expected under ISO/IEC 42001.
Still have a question? Message us on WhatsApp or book a free consultation.