1. Problem and baseline
Define the quality, production or logistics problem in one sentence, then record the current cost, frequency and handling time.
- Target process, equipment and product
- Current defect, downtime and operating baseline
- KPI and measurement method used to judge success
2. Data and interfaces
Identify when and where AI and equipment data are generated, and whether the quality is sufficient for the intended decision.
- Sample representativeness and rare defects
- Camera, sensor, PLC and MES interfaces
- Data rights, security and retention period
3. Safety and operations
Define the safe state and operating responsibility when an AI decision fails or the network becomes unavailable.
- Safety PLC, manual mode and emergency stop
- FAT/SAT, rollback and recovery procedures
- Training, SLA, patching and RMA ownership
