If an organization hasn’t been touched by AI or developed a plan to utilize it, now is the time.
Patrick Robinson, senior director at PBF Energy, said embracing AI to help streamline processes has many benefits, including making employees more efficient.
"This can be leveraged in finance, invoicing, commercial — anywhere someone is manually taking data from a form and putting it into a system to make it structured. It can be automated with AI, improve the accuracy and help with the efficiency of your employees," Robinson said. "There was a 30-year engineer who I worked with, and he was bringing up an SQL [Structured Query Language] query, which is this code on the side.
"And I was like, ‘Oh, time out. What are you doing?’ So I said, we’ll throw it into a Gen AI and make it legible, and within seconds you could actually read it. This person spent 30 years formatting queries. Think about it. You want your employees to spend more time analyzing and less time preparing and prepping all those little tasks."
Robinson was one of the panelists on the "AI Learnings from the Frontline" presentation during the 2025 AFPM Annual Meeting in San Antonio. The panelists discussed different ways in which their corporations approach and use AI technology.
CPChem, Motiva, PBF: AI adoption accelerates in petrochem sector
Eric Roberts, director of health and safety for Motiva Enterprises, said they have been looking for ways to use AI to improve safety, particularly in the area of shutdowns and turnarounds.
About a year ago, Motiva began a pilot program to see how the company could apply hazard recognition technology in the facility, an endeavor that involved multiple cameras monitoring things in real time.
Motiva decided to use it during a turnaround. The cameras were mainly clustered in the processing unit, aimed at finding weak spots and lapses in security protocol.
It took an effort to get everyone on board, Roberts said.
"This was a big ask, and we’re not unlike most of the facilities out there. We have cameras everywhere. This was going to be different," he said. "We had to make sure that all the right groups understood what we were trying to do, and the support we were going to need from them to make this happen."
It’s an ongoing learning process, Roberts said. "I would say the biggest thing we learned in going through this process was that we didn’t know how to manage this information," he said. "There’s a lot of information coming in. We wanted to put the system to the test to see how it performed for us. So we’re in the middle of doing that now."
Brent Railey, chief data and analytics officer with Chevron Phillips Chemical, said AI can be an effective tool if those using it are realistic about limitations.
"It could be extremely great technology, but … if you’re applying it across an enterprise, you’ve got to go through the scalability," Railey said. "The other side is, I’ve seen really good technology fail because the people aren’t accepting of the technology, or vice versa."
Robinson said as AI is continuing to evolve and improve, managers and employees should acknowledge that because it is a developing technology, it is not always perfect.
He gave an example of a time he asked ChatGPT how many r’s there are in the word strawberry. The technology kept giving the wrong answer — two r’s.
Robinson challenged that answer, and the technology got "a little condescending" with him. Finally, the AI admitted the mistake.
"I felt great. I was teaching a robot and I’m a ‘trust but verify’ person. These models are not perfect. They need the human in the loop for risk management, especially when you put all of your data into these systems. You need to vet that," Robinson said.
