AI planning resource
AI readiness checklist
A useful AI project begins with a specific task and evidence that the organization can support the data, integration, controls and operational change required.
Direct answer
What you should know
An organization is ready to advance an AI use case when it can define the user and outcome, access suitable information, integrate the workflow, evaluate quality, manage risk and assign ownership for operation.
Business readiness
- A defined user, task and current baseline.
- A measurable outcome rather than a generic AI objective.
- An accountable business owner and decision process.
- Representative examples of successful and unsuccessful outcomes.
Technical readiness
- Accessible, permissioned and sufficiently reliable data.
- Known systems and integration boundaries.
- Security, privacy and retention requirements.
- A feasible evaluation and monitoring approach.
Operational readiness
- People responsible for review and exceptions.
- A fallback process when the system is unavailable or uncertain.
- Training and change-management needs.
- A process for incidents, feedback and model or data changes.
Step-by-step process
-
Define the decision or task
Identify who will use the system, what they need to accomplish and the current baseline.
-
Confirm data and permissions
List the information required, its owner, quality, sensitivity, access rules and update frequency.
-
Map integrations and workflow
Identify where AI output enters the current process and which systems or people must participate.
-
Define evaluation
Create representative examples and measurable quality, safety, latency and cost criteria.
-
Classify risk
Assess impact, reversibility, affected users, privacy, security, regulation and required human oversight.
-
Assign ownership
Name business, technical, security and operational owners for launch, monitoring and incidents.