The following is a transcript of the October 19, 2026, episode of Public Power Now. Learn more about subscribing to Public Power Now at PublicPower.org/Podcasts. Some quotes may have been edited for clarity.
Paul Ciampoli
Welcome to the latest episode of Public Power Now.
I'm Paul Ciampoli, APPA's news director.
Public power utilities are increasingly looking for ways to adopt artificial intelligence as a tool to improve their operations and benefit their customers.
They are also focused on ways in which to improve load forecasting.
In early September, Santee Cooper, South Carolina's state-owned electric and water utility, announced a major digital transformation initiative.
Specifically, Santee Cooper will be utilizing Google's Weather Next 2 model to address load forecasting.
The model offers highly accurate medium-range weather forecasting.
Santee Cooper will also use Google Cloud's Gemini Enterprise to support Santee Cooper's specialized AI model.
Tami Wilson, Vice President and Chief Financial Officer at Santee Cooper, is here to provide additional details on this new initiative.
Tami, thanks for joining us on the podcast.
Tami Wilson
Thanks for having me, Paul.
Paul Ciampoli
To get our conversation started, I wanted to talk about the partnership with Google and if you could kind of elaborate for our listeners what Google brings to the table in terms of its track record and AI and weather forecasting that made it a good fit for Santee Cooper?
Tami Wilson
So we chose Google because they combined our world-class atmospheric science and AI depth with an enterprise platform around best-in-class data security and governance.
And as you know, Paul, being in public power, right, that data governance, data security is pretty important to us. So that was a key factor.
The partnership is going to leverage Google's WeatherNext 2 model.
And as you mentioned, it offers a highly accurate medium range for weather forecast, more accurate than what we think we can get with the current tools we're using.
And that capability, of course, on weather forecasting is really critical for public power utilities like Santee Cooper. And because it's the weather, we hope to get a more precise weather forecast.
It should help us from an operational efficiency and cost savings to our customers.
We evaluated whether to buy or build some of these AI tools that we are partnering with Google on.
Weather forecasting for that was kind of really a no-brainer because in order for us to build this ourselves, we would have had to have petabytes available, right?
A specialized meteorology, AI and compute, which just would not make it viable for us to do ourselves.
Paul Ciampoli
Yeah, that's interesting because obviously Santee Cooper -- it goes without saying, is a pretty large utility.
So I guess that says something in terms of your decision-making process as you looked at the options.
Tami Wilson
Right.
Paul Ciampoli
So just focusing in on the WeatherNext Two model, a couple questions for you on this area.
I'll start with the first one and give you a chance to reply.
Could you provide additional details on the WeatherNext Two model in terms of its technology and applications?
Tami Wilson
So weather, right, for Santee Cooper and utilities in general is really the single greatest variable that influences our energy demand.
This WeatherNext Two will integrate those advanced data analytics and couple it with cutting edge machine learning.
And so again, back to that highly accurate medium range forecast, the other thing it does is it's going to bring in localized weather factors.
So Santee Cooper operates two very large bodies of water, they’re lakes, Lake Marion and Lake Moultrie.
And they sit within our operating footprint. And those lakes can have a bit of a microclimate, right?
That coupled with the other things that influence our weather here in South Carolina, like humidity and storms and the coast.
And so that's one thing that was very important for us when we looked at this technology was the ability that it would bring in those localized factors.
As far as the everyday applications, we think it's going to support faster planning and more effective workflows across our generation and dispatch teams.
And so we expect that it's going to help us make better decisions with better forecasts, right, which include optimizing our generation resources -- an allocation of those resources, schedule those units more efficiently, and then hopefully respond more proactively to those changing weather patterns.
Paul Ciampoli
And second question for you on this topic.
Santee Cooper has said use of the model could yield millions of dollars in annual potential savings related to its daily decisions about serving predicted loads.
Could you elaborate on how utilizing the model could yield savings?
Tami Wilson
So as a public power entity, we are a nonprofit, as you know, and so any savings that we have goes directly to our customers.
And the way we calculated those savings, and we were conservative in our estimates, we think it's going to reduce unnecessary need for generation on certain days. It should avoid costly overprovisioning.
So obviously we are going out to meet demand if it's going to be really cold with purchased power -- power that we purchase on the market.
And so the more precise we are...it lessens the risk of us either over buying or under buying, right, which is going to make us subject to market prices.
And then the last thing I'd say in terms of that calculation was really minimize the reliance on expensive backup, peaking resources, again, because you've got more accurate weather forecasts.
Paul Ciampoli
And, you know, in terms of everything you said so far, I would imagine, I guess it goes without saying that probably sooner than later, at some point in time, this could be effectively incorporated in terms of the integrated resource planning.
Tami Wilson
Exactly.
Paul Ciampoli
So in turning to Gemini Enterprise, I want to give you an opportunity to first provide an overview of Gemini Enterprise.
Tami Wilson
Yeah, so if you think about, right, Gemini Enterprise, it's no different than what I would compare to like ERP systems that we use for our financial systems, our procurement, et cetera.
And so the fact that it's an enterprise system, it's going to allow us to use that enterprise platform, we think, to accelerate our data discovery.
And again, the big thing that from an enterprise standpoint is the data security and governance, right?
It's called tenant security, where we're a tenant on this platform.
And so it is a very secure platform and so this data doesn't necessarily go out into the internet, right, to be used by others. I think that's what that enterprise brings to us.
In addition to this, it's also the platform that's going to help us integrate a lot of our different AI modeling across our system.
So it should incorporate the WeatherNext model that we just discussed. We're also going to be using it for financial forecasting.
And obviously, we've identified other workflows that we think we can use underneath this enterprise platform.
Paul Ciampoli
And you touched on this to some degree in terms of my next question.
I wanted to give you a chance to elaborate in terms of the specific ways in which Santee Cooper is going to leverage Gemini Enterprise.
Tami Wilson
Yeah, so we think it's going to be really instrumental in automating a lot of those routine tasks that our team does every day.
It should allow our team, and we've kind of seen it in action, right, to rapidly query documentation and summarize data and it's a great way to bridge communication and data analysis across different technical silos in the organization.
So you have all that information, right, on one platform. It really should help eliminate some redundancy and again, make sure that we're consistent in the analysis that we're doing across the business.
Paul Ciampoli
Okay, could you expand on how Santee Cooper's utilization of Gemini Enterprise will benefit the utility and its customers?
Tami Wilson
There are several things that we've identified there, obviously improved grid reliability with better planning, sharper foresight, faster operational coordination.
I think the biggest thing for us is really using these tools to benefit our customers through either cost avoidance, cost mitigation, or cost savings.
And we've already seen just how it makes our team's days more efficient and quicker.
And so again, as we think about the tremendous growth that all of the utilities are experiencing across the country, we think this is going to help us really mitigate costs and be able to utilize the team that we have without a lot of overstaffing, right, which is going to be important as we think about just the constraints on the workforce, just given that demand is going up everywhere across the country.
Paul Ciampoli
If I could ask you a couple of follow-up questions just in general.
First of all, one thought occurred to me is this obviously has a myriad of benefits just for the utility and the customers as we just discussed.
Any thoughts or comments on how this is going to be communicated to the customers of Santee Cooper?
Tami Wilson
Yes, so first of all, we are not going to embark on any use of an AI tool unless it has a good return on investment and we can identify specific areas of opportunities from a cost perspective.
And I think it's going to be important to communicate that to our customers, right?
We're going through a rate study right now where we're going to be increasing rates to support the capital investment that we've got to make in our system over the next several years to meet the demand.
And one of the things through that process that they've identified is what are you guys doing to mitigate costs?
And we think this is a good tool and we think we are going to be able to show the benefits long term of implementing this and investing in this.
Paul Ciampoli
And in terms of the relationship with Google, I guess it goes without saying that this is not going to be kind of an open shut relationship in terms of, in other words, there will be an open-ended opportunity for Santee Cooper to consult with Google as these products are rolled out.
Tami Wilson
Yes, that's true, Paul.
But I will also say, right, that while we've got a great partnership with Google, they've been fantastic.
They're clearly leading edge in this space.
We do take a look at all what is available just broadly as it relates to tools. And so yes, in short, we expect that there will be other opportunities to work with Gemini.
Obviously, it is an enterprise platform, so that is a bit of an ongoing relationship.
But yeah, they've been great support right now in getting our teams, training our teams, getting them prepared for what this means for the future.
Paul Ciampoli
And one of the things I try to do with the podcast is as I speak with guests, I think a lot of times, and this is a good example probably of this scenario where a listener may want to hear more in terms of kind of like lessons learned in terms of decision making processes.
So you talked earlier about how there was a balancing or consideration of trying to do this internally versus going to an outside vendor, in this case, Google.
Could you kind of elaborate in terms of like that process and kind of any lessons learned that our listeners might derive value from.
Tami Wilson
Yeah, again....we'll start the process, right, evaluating what the right tool is and build versus buy just in terms of incorporating the organization's bandwidth, the opportunities.
You know, I talked a bit about the importance of this partnership with Google and the fact that they bring sort of this best in class data security and governance, and that's a really big thing for us.
So I would say lessons learned is be sure you have the right people at the table in the organization when you are working through trying to evaluate building something using AI, right?
Or implementing, in this case, a platform with Google, because I think it's really important that whoever the provider is, that they thoroughly understand the complexities of the business.
And you've got to make sure that those conversations are very clear in that regards.
So far, I think from the standpoint of some of the lessons that we've learned is, don't underestimate that if you make the time of what people are doing on a daily basis versus trying to implement this because while it's usually a very quick sprint, right?
So our financial forecasting project was in 8 to 12 weeks, and it's probably going to take all of the 12 weeks that we had allotted for it.
But you just can't, and this is the case with any technology, right?
You just have to be really thoughtful and build in those contingencies.
Paul Ciampoli
Yeah, and that's a great point.
And one other question occurred to me is, because this is there's an AI element to this, I would imagine and you may not be able to comment on potential uses, but I would imagine that as these platforms are utilized, there could be new ways for Santee Cooper to effectively utilize them for the benefit of the utility and its customers.
In other words, you're going into this process with a clear set of uses of these platforms, but I would imagine that there could be things that would be discovered along the way.
Tami Wilson
Yes, exactly, Paul.
We expect that's going to be the case.
And I think it's going to be, and it already is, right, where we've got to really be thoughtful about prioritizing.
And that's where we think that this, you know, we're going to be holding these use cases, you know, really taking a hard look at them about the reality of whether we can deliver on the return on that particular investment and what the cost benefits are.
You know, we established as part of our governance structure, what we're calling our innovation council.
And they did a bit of the data gathering where they went out and asked the organization and met with teams across the organization on use cases.
And I think the first pass of that, there were already about 200 use cases identified where the teams thought that AI would benefit them in their work processes.
Paul Ciampoli
That's great. So that's a great example of a public power utility kind of, proactively working to set out some possibilities in terms of use cases.
Tami Wilson
Yeah. It's pretty exciting. We're really excited about seeing some of this stuff come to fruition and be able to realize the benefits for our customers.
Paul Ciampoli
I apologize if you mentioned this, I don't think you did, but in terms of the timing of the deployment of the model and the Gemini Enterprise, are those already deployed or is there a schedule for deployment if they haven't been deployed?
Tami Wilson
They are being deployed, so I would say we're probably in the second inning here, Paul. We purchased several 100 licenses.
I think we probably deployed about 40% of those right now -- we do require the team go through training so they know how to use them.
But I think we've gotten great interest from the organization in terms of their interest in being able to utilize this technology.
Paul Ciampoli
Did you have any final thoughts or comments you want to make as it relates to this topic?
Tami Wilson
I don't think so. I appreciate the time and appreciate you inviting me to share what we're doing here at Sandy Cooper.
Paul Ciampoli
Sure thing. Well, thanks again, Tami.
And I would love to have you back at some point, perhaps a year from now, where we could revisit this topic and kind of see where things are at and how the utility and its customers have benefited from these two platforms.
Tami Wilson
I would love to do that, Paul.
Paul Ciampoli
Okay, great. Thanks, Tami.
Tami Wilson
Thank you.
Paul Ciampoli
Thanks for listening to this episode of Public Power Now, which is produced by Julio Guerrero, Graphic and Digital Designer at APPA.
I'm Paul Ciampoli, and we'll be back with more from the world of public power next week.
