Glossary term
Glossary term
Governance and Compliance
The full path from idea, intake, design, data sourcing, development, evaluation, deployment, monitoring, and retirement of an AI system. Lifecycle thinking helps governance avoid one-time approvals that go stale. Governance should attach controls and evidence to each lifecycle stage rather than relying on a single release-time review.
The OECD AI System Lifecycle framework defines six stages from Plan and Design through Operate and Monitor, widely cited in NIST AI RMF.
ISO/IEC 22989:2022 and ISO/IEC 5338:2023 define AI system lifecycle processes used to align ISO 42001 controls.
CRISP-DM and Microsoft Team Data Science Process define lifecycle stages from business understanding through deployment, applied across enterprise data science teams.