AI can illuminate strategies in energy

“With enterprise AI, it’s about having the data — 95% of our AI discussions come back to the quality, availability and veracity of data within the organization,” Whatton said.

Enterprise-level AI data needs to be centralized, or at least accessible for a centralized solution. That can be a hurdle for many energy companies. “What traditionally happens is that one plant manager optimizes Plant A, another manager optimizes Plant B and so on,” Whatton said. “That happens in midstream oil and gas, oil field services and throughout the industry — because the tradition is to look at things by asset base instead of using an enterprise base.”

This approach has led to disconnected systems that impede energy companies today. “Upstream, midstream and downstream, energy companies traditionally manage based on siloed pieces of data within separate systems,” Whatton said. “They look at the data within different teams and make decisions based on historical performance, their experience and a little bit of a gut check with what they want to do. The ones with the best instincts are usually the ones who win — and, if they hit favorable market conditions, they can win a lot.”

“AI is going to change that — but it could require a data project with an investment in technology,” Whatton said.

“Most energy companies know that technology investment is a good idea, they just hesitate to do it,” Whatton said. If energy prices are high, companies feel like their systems are sufficient. If energy prices are low, they prioritize other needs. For companies that have historically underinvested in back-office technology, the work of integrating data sources might not look like it has an immediate payoff. That commodity mindset is part of what has led to lagging tech capabilities and a resistance to change across the industry.

With so much focus on commodity prices, many energy leaders have been hesitant to prioritize or trust long-term technology investments.

“It’s less of a tech problem than it is a change-management problem for some of the senior leaders to adopt and trust new technology,” Whatton said. However, there are now two sides to that culture coin, as many current and potential employees are seeking AI technology at work. “Many energy companies are not attracting and hiring young talent that wants to be data-driven and have information aggregated where they can apply an AI model.”

Renewable-energy companies have less of a recruitment issue, but they often have less historical data for AI solutions to analyze. Oil and gas companies have rich reserves of historical data that AI can analyze for scenario planning, forecasting and more — and they need to attract the right talent. For any energy company that wants to implement enterprise-level AI, the most important step to overcoming resistance and achieving value is finding the right application.

Energy companies have a range of valuable use cases for AI, but enterprise-level value requires enterprise-level data.

Look at where your company has data that a centralized solution could access, and consider how it could inform your most important decisions or address your biggest challenges. “I would tell upstream, midstream and downstream companies to look at what their competition is doing in these spaces, adopt a technology-first integrated model and build a road map of what you need to do as an organization,” Whatton said. Leading companies are already applying AI capabilities to help inform decisions on equipment, infrastructure, utilization, strategy and other areas.

Equipment

Energy companies can use supervisory control and data acquisition (SCADA), internet of things (IoT) and other technologies to collect real-time data from pumps and other equipment. This data can inform anything from quick reactions to long-term planning.

“That type of technology can help determine if you need to perform maintenance or if things are running as they should — even within regulatory constraints,” said Grant Thornton Business Consulting Principal Jonathan Eaton. He recalled one client that regularly incurred costs because of equipment readings that weren’t precise. “They had contracts with tiered pricing based on volume, but they were in constant disagreement with customers. They didn’t have the tech to say, ‘It’s this much and this is why.’ They were constantly agreeing to settle on disputes.” Better equipment data can resolve these issues along with concerns about product leakage and other issues, while also providing the basis for enterprise-level AI forecasting, scenario building and planning.

Over time, companies can also use equipment data and demand fluctuations to identify limitations and prioritize decisions about infrastructure investment.

Infrastructure

When energy leaders need to make decisions about infrastructure, they must consider a complicated balance of factors. AI analysis and scenario planning can help them accurately understand their options, especially if they are venturing into new territory.

Infrastructure decisions at a growing renewable energy company include a different balance of considerations than those at a midstream oil and gas company that is restricted by its underground lines — those lines can force a company to prioritize supplier and customer relationships while ensuring that it is certain about any changes. AI scenario planning can help all of these companies review the possible outcomes of their investments over time, considering how a unique mix of factors might shift.

Other energy companies might need to consider dynamic factors on an ongoing basis. For instance, power companies can use AI solutions for real-time monitoring and control of their grid to improve reliability and resilience, or to manage distributed resources like solar, wind or battery storage to ensure the right capacity.

Capacity utilization

While the pipeline and storage infrastructure at oil and gas companies can be restrictive in structure, it can be flexible in utilization.

“The pipeline’s the pipeline, but that and the storage tanks can provide incredible use cases,” Eaton said. “AI can help companies better utilize existing capacity, whether it’s decisions about which trucks you’re sending where, how you’re using the pipelines or how you’re using storage.”

Grant Thornton Energy Industry Leader Tyler Jones added that companies can also view their service hours as a capacity to optimize. “How do you keep your crew as productive as possible, to drive maximum efficiency — and by virtue of efficiency, profitability?”

“Those use cases produce revenue because they get greater throughput with their infrastructure and their network,” Eaton said. “They’re using the richness of the data that’s out there to help decide what they should use, build or acquire. Companies have a big network of trucks, distribution centers and plants that move things all over the country — it just happens to be a commodity.” With the right information, companies can even consider how to incorporate entirely new commodities into their business strategies.

Business strategy

Whatton said one midstream oil company acquired a set of terminals and a truck fleet, then reworked it to take advantage of new commodity needs. “They looked at the market and adapted their truck fleet with the right infrastructure and flexible technology to start hauling water versus hauling crude oil. There’s only so much output from a pipeline, and based on where their customers were, it was a margin play at that point.”

With the right technology, processes and transition, companies can turn shifting market conditions to their advantage. Whatton cited another company that began storing olive oil. “They basically cleaned out and converted their terminals to store olive oil for import and export, because they had better flexibility on commodity pricing for that, and they could make more money.” The company moved to cloud-based terminal management technology, so that it could manage terminals in multiple locations and expand its agility to handle different products. “A lot of companies have the expertise and infrastructure to do some other interesting things, if they can get the tech right.”

As more companies take advantage of the agility that AI can provide, the industry will shift faster and in more significant ways. “You could see a second step, where midstream companies start building storage facilities to better control the pricing on the products they’re pushing through,” Whatton said. “They could be able to say they’re going to hold a product in storage, because they know what the market is going to need for capacity. I think that will be interesting, but it’s going to take the right data to predict those trends.”

“Then, it becomes about the data, and how to use the data,” Eaton said. “There’s no one silver bullet, but there are ways to overlay technology to better leverage your enterprise data.” 

“This sort of dynamic planning is going to become table stakes, because you’re going to start seeing reforecasts and reprojections monthly, or weekly, in the midterm to long-term future,” Jackson said.

Most companies start exploring AI use cases with a focus on efficiency and cost reduction. These use cases can provide a relatively familiar extension of process automations, which companies have explored for decades.

However, the expansive capabilities of AI can be applied for greater enterprise value, and that starts with recognizing the value of your enterprise data:

• Enterprise process optimization: Fine-tune enterprise-level operations in power generation, refining and distribution.
• Predictive maintenance: Forecast and avoid potential equipment failures.
• Load forecasting: Improve demand forecasting to better balance supply and demand.
• Weather adaptation: Anticipate weather patterns to optimize solar and wind farms.
• Strategic planning: Weigh market, regulatory and environmental factors with scenario modeling.
• Dynamic trading: Use algorithms to optimize trading strategies in volatile markets.
• Risk management: Identify and mitigate financial, operational and cybersecurity risks.
• Safety monitoring: Use visual analysis to help detect hazards in real time.
• Workforce training: Automate and tailor training with adaptive learning platforms.
• Customer plan personalization: Offer more tailored pricing based on usage patterns.

The key is for energy leaders to consider AI applications beyond small efficiencies and limited pilots, illuminating how data can drive business processes and decisions for enterprise priorities today and into the future.

Contacts: