Question 1

Alarm summarisation that drifts from alarm state

Accepted Answer

Operators compare the AI summary against the actual alarm list every time, for the first month. Any drift — wrong severity, wrong count, stale state — and the copilot loses the operator. The alarm state must be read from the source, never from model memory or summary.

Question 2

Asset IDs fabricated or truncated

Accepted Answer

Industrial assets have long structured IDs (site / line / cell / device). LLM text that drops a segment, or fabricates a full ID, breaks the downstream action. IDs come from the asset tool, never from model output.

Question 3

Standard-operating-procedure knowledge shoved into the system prompt

Accepted Answer

SOP knowledge ("if temperature spike in cell 3, acknowledge within 2 minutes") does not belong in the system prompt. It belongs in retrieval or in the state machine. Prompt-encoded SOPs are brittle, expensive, and fail silently on update.

Question 4

Integration with CMMS via screen scraping

Accepted Answer

Industrial CMMS systems (Maximo, SAP PM, custom) are the target the copilot must serve. Screen-scraping or CSV integration looks pragmatic and rots in a quarter. Typed integration boundaries stay stable.

AI для промышленного IoT и платформ операций