“Industrial AI has a credibility problem. I've seen these incredibly well-funded AI initiatives not deliver a result. I was speaking with one of our customers in Asia a few months back, and they spoke about how much they spent on this incredible investment, and it wasn't really delivering.”

…is clear. Meet today's energy needs while building more resilient, intelligent energy system for tomorrow. 13 conference programs, 1,800 speakers, 2,250 companies, 239,000 attendees. Over four days, Abu Dhabi became a meeting point for the global energy industry, addressing critical issues,

…to Jim Masso, who is the president and CEO of process automation at Honeywell Technologies. And I began by asking him about the latest innovations in AI, large language models and agentic AI. They are generating a lot of excitement and energy, certainly. But are they really changing the industry?

as Honeywell, you're operating in industries that are working with hazardous materials often, refining, petrochemicals, other aspects of the energy industry. A 1% failure rate there would be absolutely horrifying. So how do you think about the risks around using AI?

So I start with deterministic controls, right? We have guardrails. And the first, so we've always done things with multivariate control and optimization. This is, I think, one of the discussions around large language models changing the dynamic versus what AI has been traditionally. Yeah,

well, it's a good point, actually. Let's just dig into that for a moment. Because when we talk about AI, I think it's helpful to define terms. There seems to be an idea for some people that AI was invented in 2022 with ChatGPT. In industry, people have been talking about AI for years, maybe even…