AI asset lifecycle
The AI asset lifecycle defines the stages for managing an AI system, model, prompt, or dataset throughout its useful life.
AI asset lifecycle stages
The AI asset lifecycle consists of the following stages:
- Onboard
- The onboard stage is the introduction of an AI asset into your organization. During this stage, you can define important details about the AI asset, including the asset version and documentation.
- Assess
- The Assess stage is the evaluation of an AI asset to determine its effectiveness, its efficiency, its reliability, and its alignment with your organizational goals. This evaluation includes assessments for the performance, business and risk impact, regulatory compliance, and overall value of each AI asset.
- Build and test
- The Build and test stage is the development and testing of an AI asset to prepare it for deployment. When you're developing an AI asset, you must create the asset, code any applicable algorithms, and integrate relevant data sources. After you develop the AI asset, you can run tests to verify that it functions correctly, meets your performance standards, and produces accurate results. You can also identify and resolve bugs.
- Deploy
- The Deploy stage is the integration of an AI asset into your existing workflows. During this stage, you can also set up monitoring to track the performance of the AI asset. Two roll-out options are available: a gradual roll-out, limited to a specific subset of users, or a full roll-out, available to all users in your organization.
- Offboarding
- To retire a deployed AI asset, a user with the AI asset owner (sn_ai_asset_mgmt.ai_asset_owner) role.
For information on Completing AI lifecycle stages, see Complete AI asset lifecycle
For information on creating offboarding requests for AI assets, see Create offboarding requests for AI assets
For information on view AI assets by lifecycle stage, see View AI assets by life-cycle stage