Dive Brief:
-
Thirty-three percent of utility and energy companies worldwide have begun to pilot generative artificial intelligence — algorithms capable of generating text, images, computer code and other content — in their operations, according to a survey released last week by digital think tank Capgemini Research Institute.
-
Almost 40% of utility and energy companies have established a dedicated team and budget for generative AI, while 41% say they have taken a “watch and wait” approach to the technology. But 95% of utilities and energy companies said they have discussed the use of generative AI in the past year, according to the survey.
-
A third of survey respondents in the energy sector said they are testing the ability of generative AI to create realistic datasets that can be used to shorten development timelines, but early AI adopters in the energy sector say the industry has only begun to experiment with the AI use cases with the greatest potential.
Dive Insight:
Utilities are generally conservative in their adoption of new technologies, but they seem to be keeping pace with most other industries when it comes to the rise of generative AI, according to Doug Ross, vice president of data and insights for Capgemini.
Energy and utility companies that participated in Capgemini’s survey were on pace in their adoption of generative AI with other industries. Thirty-nine percent of energy sector companies have dedicated teams and budgets for generative AI, compared to a global, all-industry average of 40%, according to the survey results.
Although ChatGPT intensified public awareness of generative AI in recent months, the technology itself has been around in various forms for at least three years, Ross said. Even so, generative AI has attracted growing interest from a variety of industries — and Capgemini’s survey data suggests utilities are no exception.
Utilities tend to see generative AI as having the potential to accelerate growth, rather than posing a potential disruptive threat, Ross said. And their stated assumptions align with his experience with the technology. He cited a case study of an insurance company that planned to implement AI in its call center to reduce the average call time with customers. Call time did not go down, but the company saw overall sales increase. Ross said he believes this is because the AI reduced the amount of time spent on rote tasks like information collection, which allowed the customer service representatives to spend more time developing relationships.
Fifty-two percent of energy companies indicated an interest in deploying AI on their sales teams in Capgemini’s survey. They also showed interest in more technical uses of generative AI, such as using it to generate realistic but synthetic data to support IT and development processes.
But the focus on using AI in these capacities means most utilities are still experimenting with AI at the “surface level,” said Raj Chudgar, a consultant for data center provider EdgeConneX.
EdgeConneX began piloting an AI service from energy supplier Gridmatic in January to test whether artificial intelligence could help the company achieve the 24-7 clean energy standard touted by Google. The company had been using annual renewable energy credits to offset its energy use for a couple of years, but wanted to take their sustainability goals to the next level, said Anand Ramesh, senior vice president of advanced technology of EdgeConneX.
The company’s initial goal was to match 80% of their electrical use with hourly clean energy by the end of the two year period without significantly increasing their energy costs. The AI has nearly achieved that goal within the first few months, and should approach 90% by the end of the year, exceeding expectations, Chudgar said. However, he said it is unlikely the AI will be able to achieve 100% renewable energy given the current resources on the grid without incurring significant costs.
Supplying 24-7 clean energy is just one of three AI use cases that has emerged since Gridmatic began working with artificial intelligence six years ago to optimize bidding in wholesale energy markets, Leesa Lee, chief marketing officer for Gridmatic, said. The company has also seen success using AI to optimize the operation of energy storage assets, and to help manage demand-side efficiency programs. While there is a place for AI chatbots and the like in customer service, AI’s greatest potential benefits will be realized elsewhere, Lee said.
“The customer-facing, front-line impacts will probably be much more apparent and much more immediate,” Lee said. “But there will be deeper things that will potentially be more hidden, and will have greater impact.”
Ross agreed that there will likely be two sets of use cases for AI in the utility sector — more immediate, low risk possibilities like using AI to generate posts on social media, and higher-risk functions related to a utility’s core activities like grid planning. The latter, Ross said, will likely take longer to implement if only because utilities will have to get regulators to sign off on these uses of AI.
And the full implementation of this latter group of AI use cases could take much longer than many expect — perhaps 10 to 20 years, Chudgar said. Private commercial entities like EdgeConneX may be able to move a little faster because they don’t face the same regulatory hurdles, he said. But even for the competitive markets, number of skilled professionals who can build and deploy AI is limited, according to Chudgar.
