Revise requirements using technician feedback before: What should happen next?
Technicians identified a safety-relevant requirement, so revise the design before development proceeds.
The question
An energy company is designing an AI system to prioritize maintenance alerts. During stakeholder interviews, technicians report that the proposed alerts omit equipment conditions they use to judge immediate danger. The project team believes the model can still meet its original accuracy target. What should happen next?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Add more historical maintenance records without changing requirements.More data may improve prediction but cannot ensure the system represents technicians’ identified safety information needs.
- Continue development and train technicians on the alert format.Training does not resolve missing safety-relevant requirements identified by people with operational knowledge of equipment conditions.
- Deploy under technician supervision and collect feedback afterward.Post-deployment feedback delays addressing a known safety gap and may expose workers or equipment to foreseeable harm.
- Revise requirements using technician feedback before finalizing the design. ✓Stakeholder feedback reveals a material safety requirement, so the team should update intended use and alert content before proceeding.
The trap
Treat credible stakeholder evidence about omitted harms or needs as a requirements change, not merely post-launch feedback. How to remember it
Technicians identified a safety-relevant requirement, so revise the design before development proceeds.
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
- Define unsafe-weather override triggers and a fallback: What control should be completed before deployment? →
- Use the smaller model: Which choice is most defensible? →
- Test misuse scenarios and constrain unsupported inferences: What design decision best addresses the risk? →
- 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.
