Brian Taylor, General Manager for Tennessee public power utility CDE Lightband, and Lance Haynie, Government Affairs Director for the City of Santa Clara, Utah, on June 10 provided a variety of examples of how artificial intelligence can be utilized by public power utilities and public power cities.
They participated in a panel at APPA’s National Conference in New Orleans, La., “Practical Applications of AI for Utilities.”
Taylor said that the first thing his utility did was create an AI policy. “We’ve locked in now on using Google Gemini as our platform,” he noted.
With the policy, “anything that you’re wanting to do with AI, you’ve got to prove to us that it’s mission critical or it goes to the core of our business.”
CDE Lightband built a dashboard “and you can only access our AI programs through that dashboard.” At the same time, Taylor noted that the utility’s communications and marketing teams “have full access to AI where they can do things to be a little bit creative.”
In one of the slides for his presentation, Taylor listed the following ways in which utilities can utilize AI:
• Knowledge Retrieval
• Customer Understanding (sentiment)
• Customer Segmentation
• Anomaly Detection in Data
• Satellite Imagery Analysis
• Power Demand Forecasting
CDE Lightband has built a centralized data warehouse as part of its overall AI-related activities.
In providing examples of how AI can be utilized, Taylor noted that “I’m constantly asking someone to look up” things like “when did the board approve our new vacation policy? When did we approve the Silicon Ranch contract?”
CDE Lightband is “scanning in all of our board minutes and then you can” utilize AI to research prior board action and the result is that “it will give you the date that it was approved, and actually how it was voted on, what board members voted on it as well.”
The utility is going all the way back to 1938 to scan board documents and make them available through AI.
“That’s an easy thing to do, but it’s amazing how much time and money it’s already saved us.”
Another example of how CDE Lightband is using AI involves advanced metering infrastructure.
“Most everybody has an AMI system now – you’ve got all this data on meter reading and such. We are also in the broadband business so there’s things we’re wanting to do to target market those folks,” he said.
“What we’re doing now is taking all usage data from the AMI and then clustering those into patterns and like customers so now we’re segmenting customers into like usage patterns,” Taylor noted.
CDE Lightband recently rolled out a smart thermostat program. “Rather than targeting 86,000 customers, there may be 20,000 customers that we think…this is ideal for them. So now we’re just targeting those 20,000 customers.”
For his part, Haynie said that AI “improves efficiency consistently…if you have a small staff, they’re able to get a lot more done with their time. It really does help and it does enhance decision making.”
Haynie said that “we train it on what our decisions are” including, for example, information related to legislation.
Haynie also touched upon a “Human in the Loop” philosophy when it comes to AI. “For us, we have a philosophy of AI can’t do anything on its own.” Human in the Loop is making sure that “a person is constantly involved.”
Load Forecasting
Haynie detailed how AI has helped the city when it comes to load forecasting.
“We have hit 92 percent accuracy. Now, we are in Utah, where it’s sunny 300 days out of the year, so our weather is a little more predictable than some other regions,” Haynie said.
He noted that “we don’t buy all of our power…so our forecasting is about what is the demand going to be and can we generate it in house or do we have to go out in the market and buy the excess.”
Haynie said “it’s a dollars and cents thing at the end of the day. We went from” around $15,000 in savings “to a quarter million in one month. It was huge when we saw that.”
That is because “we went from manual spread sheets – we had a scheduler who was inputting weather data from weather.com…pulling in load data from a bunch of different sources,” he said.
“Our system now goes back and it pulls all of that directly…and it’s iterative. It’s not just an algorithm that’s running the math behind it and spitting out a number. If it messes up it actually knows that it messed up, sends us an alert and then tries to fix it all by itself.”
In one of the slides for his presentation, Haynie also details how AI can be used for administrative efficiency, listing the following:
• Meeting Minute Summarization: Using AI to condense lengthy meeting minutes into concise and actionable summaries.
• Document Analysis: Using AI to extract key information from lengthy documents, reports, and emails.
• Time Savings: AI applications can reduce hours of administrative work.
• Implementation Ease: Tools like ChatGPT are low hanging fruit requiring minimal technical expertise.
