From its beginning in simply reporting when a pump shuts down, to now predicting that event before it happens, information technology has greatly evolved in the last 10 years—and that evolution is only getting faster. No longer is it enough to have the data—it’s necessary to maximize it. Operations like artificial intelligence, machine learning, data integration, and the like are now the soup du jour. The three participants in this story were unified on the fact that maximizing the computing power of AI exponentially benefits planning, modeling, profitability, and human safety, among other things.
For Caterpillar’s Andy Publes, IT/Automation’s secret sauce is in collaboration.
Experts like Publes know that consistency of operational optimization is among the great achievements of automation—a consistency applied (and informed by) the firehose of data that streams in from today’s digital oil field. But there’s a human element behind all that, which is a system’s “secret sauce” according to Publes. Publes is the drilling and completions segment manager for Caterpillar Oil and Gas—a branch of the main Caterpillar brand.
Connecting the knowledge of longtime system developers with the extensive field understanding of end users is an IT benefit that’s not often noticed, he said. But those connections create conversations in which “the engineer is now able to put context into some of the data they have in the field, and the operators are also looking at the engineering and learning insights on what they can do differently with the equipment.”
The human connections actually start from the beginning of the company’s association with a client.
“We have learned,” Publes said, “that in order to be successful with these tools we want to create this ‘customer-in-the-loop’ scenario in which you create a framework where you continue the development of the customer.” Including the customer at the ground floor speeds the process and helps developers make sure they don’t overlook issues that may be unique to the customer.
“At the start, when you have a system replicating tasks [formerly done by a human, there are things that you don’t know that you don’t know. So by including the customer from the start, you help yourself develop solutions for those nuances,” he said.
How That Works in the Field
“Optimization” in automation and IT sounds great. So in the real world, how does that play out? For Publes, something as seemingly simple as monitoring a piece of equipment’s operation for the purpose of shutting it down when it’s idle can save 10-15 percent of fuel costs. This is significant because fuel is among the biggest expenses in well servicing.
What’s not so simple about deciding when equipment is idle enough to shut down is that it’s more than just whether it’s been idling for a certain amount of time, although that is part of the equation.
“The computer that is onboard evaluates patterns of how the system is being used” he said, including, “Is there anything else going on outside of that component that may indicate that the system is about to be used?” With that information component in play it’s not going to be shut off only to be restarted one minute later. Also part of the shutoff decision is the amount of time the machine requires to start up and reach operating speed.
All this involves receiving and analyzing thousands of data points, across an entire operation, something no human onsite has time to do. Yet the payoff is the significant savings in fuel costs and in ESG-related emissions.
The Advantage for Humans
Publes, like others in the technology sector, stresses that automation does not replace people. “You still need an operator,” he said, adding, “It’s just that the type of work the operator is doing has changed.”
He continued, “There is a factor of freeing that person—not completely—but freeing them to be performing other jobs.” In the process, “they can be useful elsewhere in the organization.”
Safety is also a top priority, as machines reduce the need for employees to put themselves in dangerous situations.
Cognite VP: AI Lets IT Boost Profits and Safety
In its day, just the ability to collect and analyze data was a huge step forward. Now, says Cognite Global VP, Partner Success Prabu Parthasarathy, there’s much more to be gained from that knowledge when it informs artificial intelligence (AI) and especially generative AI.
“Benefits can range from operational efficiency improvements to reduced emissions to improved maintenance and reliability,” he said in an email interview. “The power of generative AI combined with contextualized data helps provide faster access to complex data through natural language queries.”
In practical terms, Parthasarathy noted that combining Gen AI with contextualized data lets operators build “faster and more complex optimization models that provide real-time feedback,” which allows faster problem solving and improved operating efficiency.
He stressed that this is an ongoing process because the AI space is fast-moving, “and use cases are evolving rapidly.” This responsibility doesn’t just fall on IT staff, he said, but must include ongoing buy-in from the executive suite to field operations, as well as from IT staff.
Analysis Avoids Paralysis
AI improves on simple IT because it “can deliver accurate product forecasts, integrate diverse data sources, monitor equipment health in real-time, and optimize operational decisions based on need-based metrics, like profitability,” he said. And Gen AI moves the needle further yet by “streamlining and expediting tasks and retrieving necessary data.”
As noted by Publes, safety is a significant feature. One example, said Parthasarathy, is by informing drones used for inspecting dangerous areas, keeping humans out of harm’s way.
There is also the predictive maintenance side, where AI lets operators be proactive on maintenance schedules, reducing equipment downtime. This has the double benefit of lowering expenses by preventing more costly repairs, along with boosting revenue by reducing loss of revenue when pumps aren’t pumping oil.
Merging Companies and Data
M&A activity hit a record high in 2023, and combining those siloed databases can be challenging—but the rewards are significant. For Parthasarathy, AI can greatly speed the process and unearth valuable, actionable information. “This is where a system that brings together data and contextualizes them becomes crucial to providing key insights to relevant executives. AI helps unearth key insights that would help optimize the portfolio to maximize profitability in these companies, drive the requisite cost reduction/synergy savings, and meet market expectations.”
Merging Internal Siloes
Parthasarathy is also on the “no more siloes” track with AI. “We’ll start to see a more holistic approach to digital solution development,” he said. “For example, an oil and gas company may have one team
looking into autonomous operations while another team focuses on real-time simulation modeling, and a third tries to optimize alert management at the same time. These projects are actually connected use cases, and by solving the industrial data and AI problem with liberated data, they can build on one another to achieve more efficient success.”
IT Safety Includes Data and People
It’s much like inspecting the area around a car or service truck before jumping in and taking off, said Danos’s Sonny Orgeron, the company’s director of information systems and security. Before using AI to inform decisions based on cloud databases, staff members should follow company data safety protocols, he pointed out, although the dangers may be different. “Digital safety might not cause someone to lose a life or limb, but it definitely can cripple a company’s reputation and its ability to function as a business, to service customers and employees,” he noted.
Acknowledging that the vast majority of data breaches are caused by an employee failing to follow existing security protocols, Orgeron has a recommendation. If every application an employee wants to install has to be approved by the data governance team or information services governance team, it would help. “You put in some trip bars, for lack of a better term, and if an application trips on this, it’s going to keep someone from implementing it. It has to constantly evolve because there are so many applications, technologies, and joint technologies that come out.”
Orgeron’s phrase for the ongoing diligence is that “It’s not like a George Foreman—you can’t set it and forget it.”
On the more human side, he said Danos has implemented procedures for its service vehicles that do involve the 360 degree pre-driving walk-around before driving off, and it also includes software in the vehicle that “helps them be mindful and focus on driving safety,” he said.
It’s a Helper, Not a Decision Maker
Publes, along with others, has pointed out that AI can and does sometimes make mistakes in conclusions, just like humans do. So while it can greatly speed data collection and analysis, humans cannot take all decisions as gospel. The future of AI may lie in its becoming more accurate as it gathers more data, but, at least for the foreseeable future, humans are still necessary.
Paul Wiseman writes in the oil and gas sector. His email address is fittoprint414@gmail.com.