An AI use case can change even when its name and business purpose stay the same. A new model version, prompt, retrieval source, threshold, vendor feature, tool permission, or data source can alter behavior enough to make the previous review incomplete.
Define change categories in advance
List the changes that require no review, a focused review, or a full reassessment. Consider model or provider changes, new data, expanded users, new automated actions, different decision authority, new integrations, and material changes to monitoring results.
Compare the change with the approved use case
Ask whether the purpose, affected people, outputs, data, control environment, or downstream consequences have changed. A technically small update may still be significant if it changes what the system is allowed to do.
Reuse evidence carefully
Prior tests and approvals can remain relevant, but identify which evidence still applies and which assumptions need to be retested. Record the rationale rather than treating an old review as automatically portable.
Update the inventory and owners
Reflect the new version, date, dependencies, review result, and next trigger in the use-case inventory. Confirm that control owners and monitoring remain appropriate.
Keep regulated determinations separate
A model change review does not decide legal permissibility, clinical appropriateness, financial suitability, employment fairness, safety, privacy compliance, or other professional requirements.
Related knowledge
Use this with AI Use-Case Inventories, AI Risk Acceptance Records, AI Governance Review Triggers, and AI Governance Control Owners.
References
This resource supports operational governance. It does not establish legal compliance or replace qualified legal, regulatory, clinical, financial, employment, safety, privacy, accessibility, or other professional review.