The first ASQ Raleigh Section open discussion meeting of 2026 was well attended, with 14 participants, several of whom were newcomers! Following the lively discussion on AI in quality at our December meeting, Chris Andrassy volunteered to lead a session on integrating AI into nonconformance investigations and reporting. We discussed his implementation in which data is collected, guided by AI, followed by AI analysis, suggested likely root causes, and report generation.

There was extensive discussion about the interaction and roles of humans and AI. While AI can provide consistency, efficiency, and sometimes insights, it can also make mistakes and requires human oversight. Is there an issue with “oversight fatigue”, where human experts don’t notice AI errors? One suggestion was to use the interaction between the AI agent and the human, in which the AI agent asks the human probing questions.

It seemed as though a similar capability could be created for change control.

It was noted that, as with other quality areas, a process needs to be established with valid data before automating it with AI. To analyze the problem and identify potential root causes, AI needs a sufficient volume of accurate historical data.

Tim Whetten