AI testing for industry

If your technicians query manuals and procedures through an assistant, the test works like a factory acceptance test, with criteria agreed before it starts and questions from a normal shift and from a breakdown.


AI testing for industry. If your technicians query manuals and procedures through an assistant, the test works like a factory acceptance test, with criteria agreed before it starts and questions from a normal shift and from a breakdown.

What can go wrong

An invented example of how it can fail: A shift technician asks for the tightening torque of the bolts on a flange on line 3. The assistant gives the value from the manual of another pump model. The value comes from a real document, so it looks reliable. The error shows weeks later as a leak, and the incident investigation traces it back to the assistant’s answer.

Questions the test would include:

What the AI Act means here

An assistant that looks up manuals and shift reports is generally not high-risk. Since the 2026 amendment, a system used solely for non-safety-related aspects of user assistance does not qualify as a safety component (Article 6(1a)). It does if its failure would endanger health and safety (Article 6(1b)). AI in machinery is now handled within Regulation (EU) 2023/1230 on machinery, which has moved to Section B of Annex I.

In any case, if the assistant talks with people, they must be told they are talking to an AI unless that is obvious (Article 50, from 2 August 2026). And whoever develops it or uses it in their organisation must take measures to support the AI literacy of the staff who work with it (Article 4). Whether a system is high-risk is a legal question for your adviser. I do not classify it; the test gives you evidence either way.

The test follows the acceptance logic your plant already knows, with criteria agreed beforehand and a test bench your team can repeat in its own environment.