Let’s talk about the real AI applications in industry
What I think when the topic is “new generation systems”: AI.
AI applications in industry are showing 10% to 99% improvement. This is not from using Chat GPT!
We are talking about the wrong things
I was part of a panel discussion last week at ASQ’s RMMR conference. The topic: “Safety and Reliability Implications in New Generation Systems and Processes.”
Leading up to the panel and the day before, I would try to approach people to talk about AI. Mostly, they jumped to Chat GPT, proprietary lockdowns, and then didn’t want to talk much more about it. I thought that was curious.
Some showed up early to attend a pre-conference course about data mining and machine learning. Students used JMP for practical applications. From what I heard, it was a success!
Data mining and neural networks are not the AI I think about. Related? Yes…
What we should be talking about
When I talk about AI in new generation systems, I’m talking about AI Agents, Agentic AI, Generative AI and LLMs used in the manufacturing industry. And, yes, used to improve safety and reliability.
Co-panelist, Julio Lebrón-Feliciano, approached the topic from critical systems safety and reliability. And co-panelist, Peter Curtis, from integrated systems reliability and worker safety.
All of these are areas where we can benefit from using AI.
In the case studies I presented, there was no less than 10% improvement in areas that used AI Agents. Up to 50% improvement when looking at AI Agents integrated with Generative AI/LLMs. And 99% improvement when creating systems that use AI Agents, Agentic AI, Generative AI, and LLMs together.
Improvements are based on the process. No matter what, 10% is nothing to sneeze at. And 99% is practically unheard of.
Is it all roses? No. This is new tech involving autonomous decisions and can be complex. I interviewed Jake McKee, a thought leader on AIX, whose advice was to SLOW DOWN. I agree.
Adopt new tech slowly
Don’t go gangbusters and implement everything all at once. Take it slow. Evaluate if your process can benefit from AI. Start with AI Agents, and don’t give them all the decision-making capability. Use AI to help with all that data. Use AI to add the human-in-the-loop for decisions. Start slowly. Learn about what works and what doesn’t.
I think it would be a mistake to limit our conversations about AI to Chat GPT, Grok, or Gemini. Even if in passing.
Want to see how AI Agents are driving these results in manufacturing? My ASQ RMMR presentation includes real-world case studies and valuable references. Comment ‘AI Agents’ below, and I’ll send you the link directly once it’s publicly available!
I talk about more than AI. If you’re an engineer in industry, join me: newsletter.deeneyenterprises.com