A team wants to know as soon as a deployed model's behaviour meaningfully changes. Which practice gives them that visibility?
Select an answer to reveal the explanation.
Short Explanation
If you need to know as soon as a deployed model's behaviour changes, you need continuous monitoring of data and model drift. That's the only practice here that surfaces changes as they happen.
Full Explanation
The correct answer is the practice of monitoring data and model drift on an ongoing basis after deployment, so changes in behaviour are surfaced as they occur. This is the only option that produces a signal at the time behaviour actually changes, rather than long after the fact or indirectly.
The option describing re-validating the model once a year during the annual governance review is incorrect because an annual review would only catch a change up to a year after it happened.
The option describing reviewing customer support tickets for complaints about the model is incorrect because support tickets are an indirect, delayed, and incomplete signal that depends on users noticing and reporting a problem.
The option describing comparing the model's documentation to its original design specification is incorrect because documentation does not change on its own and comparing it to a fixed specification reveals nothing about live behaviour.
For the exam, when the question asks for timely detection of behaviour change, the answer is ongoing drift monitoring.