Stage 3 – Deploy
Stage 3 of the AI Lifecycle transitions a capability from build to operational use. This stage typically lasts 2–8 weeks and focuses on making the governance and change machinery fully operational before the capability touches production data.
Key activities in Deploy:
- Delegation-Authority Register entry (GOV-14)
- Register the capability as Tier 3+ (where applicable).
- Record accountable executives, responsible owners, and delegated authorities for decisions, exceptions, and incident response.
- Materiality Calibration population (GOV-16)
- Calibrate the capability’s materiality / criticality based on impact, scale, and sensitivity of use.
- Confirm tiering, required controls, and monitoring intensity.
- Evidence Register row population (GOV-13)
- Create and populate the Evidence Register entry for the capability.
- Link to design artefacts, testing results, risk assessments, and control evidence that demonstrate compliance with policy and standards.
- User training delivery (CHG-01 Step 5)
- Deliver training to end users, support teams, and control owners.
- Cover safe and intended use, limitations, escalation paths, and incident / issue reporting.
- Rollout per the change management plan
- Execute the agreed rollout strategy (e.g., pilot, phased rollout, or big-bang) under CHG-01.
- Monitor adoption, performance, and early risk signals; adjust rollout steps as needed.