ResourcesA clearer starting point.
Practical resources.
Use these short checklists to prepare a discovery session or review the readiness of an existing initiative.
Before selecting a modelAI opportunity brief
- Name the user and the decision or task.
- Describe the current process and its cost or limitation.
- Identify trusted sources and access restrictions.
- Define success, important errors and a non-AI baseline.
- Name a sponsor, an operational owner and the next decision.
Before widening a pilotEvaluation planning checklist
- Create representative examples and difficult edge cases.
- Separate development examples from final evaluation.
- Define acceptable errors and mandatory refusal conditions.
- Record versions of models, prompts, data and tools.
- Agree review gates, rollback and operational monitoring.
Before committing to deliveryData readiness questions
- Who owns the source and approves its use?
- Is the data representative of the intended workload?
- How are labels, missing data and duplicates handled?
- What freshness and lineage do consumers need?
- Who responds when quality checks fail?
These checklists structure a discussion; they do not replace a project-specific assessment.
