Templates › AI initiative business case
AI initiative business case
Template id: business-case Responsible skill: opportunity-management:business-case
A proportionate case for one significant AI initiative. Choose the size first: for a light case, complete sections 1 to 6 and 9 briefly; for a full case, complete everything. The programme-level funding case uses the investment-case template instead. Draw on existing records (the opportunity-canvas, the value hypothesis and the effort assessment) rather than writing new content. Anything not yet known stays marked [To confirm: ...].
1. Decision requested
| Field | Value |
|---|
| Opportunity reference and backlog stage | [To confirm: reference and stage] |
| Case size (light or full) and why | [To confirm: size and reason, for example cost, risk tier, reach or cross-unit funding] |
| Decision requested, of whom and by when | [To confirm: one paragraph] |
| What happens if the answer is no | [To confirm: consequence] |
| Business owner presenting the case (by role) | [To confirm: role] |
2. Problem and outcome
| Field | Value |
|---|
| Framed problem statement | [To confirm: from the canvas] |
| Affected roles and approximate numbers | [To confirm: roles and numbers] |
| Strategic outcome supported | [To confirm: outcome or "none"] |
| Risk tier and reviewer gates required | [To confirm: tier and gates] |
3. Options
At least three, including doing nothing and a non-AI or lower-rung option.
| Option | Description | Expected value (range) | Cost (build, assure, run) | Main risks |
|---|
| Do nothing | [To confirm: what continues] | [To confirm: none or range] | [To confirm: cost of doing nothing] | [To confirm: risks] |
| Non-AI or lower-rung option | [To confirm: process change, automation, training or adoption scenario] | [To confirm: range] | [To confirm: costs] | [To confirm: risks] |
| Recommended AI option | [To confirm: pattern and scope] | [To confirm: range] | [To confirm: costs] | [To confirm: risks] |
Why the recommended option beats the non-AI option: [To confirm: reasoning]
4. Costs of the recommended option
For a stated period. Licence and consumption figures come from the organisation's agreements or a licensing partner.
| Cost | Range | Confidence | Owner (role) |
|---|
| Build (design, configuration or code, knowledge preparation, integration) | [To confirm: range] | [To confirm: high, medium or low] | [To confirm: role] |
| Assure (testing, evaluation, reviews) | [To confirm: range] | [To confirm: confidence] | [To confirm: role] |
| Run per year (owner time, knowledge upkeep, review cycles, support, consumption) | [To confirm: range] | [To confirm: confidence] | [To confirm: role] |
| Adoption (learning, communications, Champions support) | [To confirm: range] | [To confirm: confidence] | [To confirm: role] |
| Retirement | [To confirm: range] | [To confirm: confidence] | [To confirm: role] |
5. Benefits
From the agreed value hypothesis only.
| Field | Value |
|---|
| Mechanism of change | [To confirm: mechanism] |
| Primary measure and baseline (with date and method) | [To confirm: measure and baseline] |
| Expected range and timing | [To confirm: low to high, and by when] |
| Cashable and non-cashable split, agreed with finance | [To confirm: split and finance role] |
| Position at the low end of the range | [To confirm: summary] |
| Benefit owner (role) | [To confirm: role] |
6. Risks
| Risk | Likelihood and effect | Mitigation | Owner (role) |
|---|
| Delivery | [To confirm: assessment] | [To confirm: mitigation] | [To confirm: role] |
| Adoption | [To confirm: assessment] | [To confirm: mitigation] | [To confirm: role] |
| Knowledge or data readiness | [To confirm: assessment] | [To confirm: mitigation] | [To confirm: role] |
| Security and privacy (per risk tier) | [To confirm: assessment] | [To confirm: mitigation] | [To confirm: role] |
| Responsible AI (tier 3 and 4: impact assessment and oversight) | [To confirm: assessment] | [To confirm: mitigation] | [To confirm: role] |
7. Evaluation and quality (full case)
| Field | Value |
|---|
| How quality will be tested before release | [To confirm: approach] |
| Acceptance threshold agreed with the business owner | [To confirm: threshold] |
| How quality is re-checked after changes | [To confirm: approach] |
8. Sensitivity (full case)
| Scenario | Effect on the case |
|---|
| Adoption at half the planned level | [To confirm: effect] |
| Run cost at the top of the range | [To confirm: effect] |
| Benefits at the low end of the range | [To confirm: effect] |
9. Exit points and decision
| Checkpoint | Date | Result that would stop or redesign the initiative |
|---|
| [To confirm: for example end of pilot] | [To confirm: date] | [To confirm: disproof condition from the value hypothesis] |
| [To confirm: for example six months after release] | [To confirm: date] | [To confirm: condition] |
| Field | Value |
|---|
| Reviewer findings (blocking or advisory) | [To confirm: findings, or "none required at this tier"] |
| Finance review (role and date) | [To confirm: role and date] |
| Decision and conditions | [To confirm: decision and conditions] |
| Decided by and date | [To confirm: body and date] |
Credit: Template from CopilotForge by Simon Foote (https://github.com/SimonFoote-greensparx/CopilotForge), licensed under CC BY 4.0. Keep this credit when you share or adapt the template; in your own completed document you may keep it as a footer or remove it.
Download the template (Markdown)