Continuously monitor performance and drift metrics: Which approach best satisfies both the ongoing-monitoring
Continuous drift monitoring combined with a defined cadence and threshold-based triggers meets both the monitoring and scheduled-maintenance requirements.
The question
A deployed demand-forecasting model performs well at launch, but the operations team knows customer behavior shifts seasonally and after promotions. Governance requires continuous monitoring plus a regular maintenance and retraining schedule. Which approach best satisfies both the ongoing-monitoring and the scheduled-maintenance requirements?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Run a thorough one-time validation at launch and freeze the model, since a system proven accurate in testing will remain reliable across the seasonal cycles.Freezing a model ignores drift from seasonal and promotional shifts; launch validation cannot guarantee ongoing reliability, and there is no monitoring.
- Continuously monitor performance and drift metrics, and set a defined retraining and maintenance cadence with drift thresholds that can trigger earlier updates. ✓Combining live drift monitoring with a scheduled cadence, plus thresholds that can trigger off-cycle retraining, satisfies both continuous monitoring and regular maintenance.
- Wait for users to file accuracy complaints and retrain only when volume warrants it, since complaint-driven upkeep concentrates effort where it demonstrably matters.Complaint-driven upkeep is reactive and lacks a monitoring signal or a regular schedule, so degradation can persist unmeasured between complaints.
- Retrain on a fixed annual date regardless of behavior, since a predictable yearly cadence keeps the maintenance schedule simple and easy for auditors to verify.A rigid annual retrain has a schedule but no continuous monitoring, so seasonal and promotional drift within the year goes undetected and unaddressed.
The trap
Providing a fixed maintenance cadence while omitting the continuous monitoring that detects drift between scheduled updates. How to remember it
Continuous drift monitoring combined with a defined cadence and threshold-based triggers meets both the monitoring and scheduled-maintenance requirements.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding How to Govern AI Development questions
- Complete the applicable conformity assessment and prepare: Which combination is the correct precondition →
- Red teaming that has skilled testers adversarially attempt: Which periodic activity most directly probes that →
- Triage and contain the incident: Which response best meets both the management and the documentation →
- All 426 Understanding How to Govern AI Development questions →
Part of the Certsqill AIGP question bank · Understanding How to Govern AI Development ·
Every answer, right and wrong, comes with its own explanation.
