Enterprise AI — Course 23: Operating and Changing AI Services

Lesson list

1. Live signals and respondersThe learner should be able to interpret technical, task, useful outcome and human workload evidence, identify its limits and connect it to proportionate authorised action.
2. Version and dependency recordsThe learner should be able to reconstruct the effective service combination, distinguish identity evidence from causal proof, and repair only the traceability gaps relevant to the decision.
3. Release and controlled rolloutThe learner should be able to judge evidence against affected behaviours, apply declared exposure rules and choose a supported scope within release authority and recovery conditions.
4. Incident triage and containmentThe learner should be able to distinguish confirmed impact from hypotheses and unresolved scope, then choose authorised containment and a usable evidence handover.
5. Recovery and restorationThe learner should be able to separate a recovered configuration from external effects, reconcile uncertain outcomes, and judge restoration against the supplied conditions.
6. Users, support and operational learningThe learner should be able to interpret correction effort, non-use and complaint coverage, then connect an operational improvement to an accountable action and observable result.
7. Retirement and dependency exitRetirement decisions follow actual unfinished work, consumers, access and record responsibilities, with evidence for each declared exit condition.
8. Operating through a material changeThe learner should be able to use a tested improvement, the decisive changed dependency, explicit operating conditions and stated ownership to choose a revisable service disposition.
9. Operating decision assessmentEight required reasoning groups assess operating decisions using new self-contained cases. One of three variants is served in each group; every group is required.