# Ai Solution Canvas

## 1. Problem and value
- Problem:
- User:
- Desired outcome:
- Value created:

## 2. Inputs and data
- Inputs available:
- Data source and owner:
- Quality and coverage:
- Consent, rights and access:

## 3. Ai capability
- Perception, prediction, reasoning, generation, decision support or action?
- Baseline method:
- Proposed model, algorithm or service:
- Why this approach is suitable:

## 4. Output and workflow
- Output format:
- Person or system receiving it:
- Action following the output:
- Human review or approval:

## 5. Evaluation
- Technical metrics:
- User or business metric:
- Important groups and conditions:
- Failure cases:
- Acceptance threshold:

## 6. Responsible use
- Privacy risk:
- Fairness risk:
- Security or misuse risk:
- Transparency requirement:
- Escalation and appeal:

## 7. Production
- Deployment environment:
- Latency and availability needs:
- Monitoring signals:
- Retraining or update trigger:
- Operational owner:

## Final decision
- Build / pilot / revise / do not proceed:
- Evidence needed next: