Live from DTECH: How Cross-Functional Data Powers Modern Utilities

Alright. Hi, everyone. My name is Christina Park. I'm the senior director of energy strategy at Skydio. Skydio is the leading autonomous drone manufacturer. We are actually based in San Mateo, California, and we're very excited to be here with UIGA as well as D Tech.

A little bit of background for me. I've been at Skydio for three years. I was hired actually when Skydio decided to verticalize and focus on very specific industries, one of them being utilities. And they were chosen based on the critical industries that society relies on.

So using autonomous drones and remote operations to be able to improve your reliability and resilience, that's the mission of our company. But I was actually hired in because I was once upon a time a customer. So I spent fifteen years of my career working at the New York Power Authority. I was in engineering and asset management and happened to come across Skydio as we were investigating their offerings in terms of how to operationalize technology in asset management.

So very excited to lead this discussion today, and I will hand it over to AP.

Great.

Thank you. Good morning, everyone. My name is AP. I'm from Exelon. I'm the director of customer strategy.

Been with the company for little over sixteen years. Been in various operational departments, planning, launched a drone program, was the head of innovation and investment strategy, and now in the customer side. Really excited to be here, especially, you know, because of the different departments I've been in. This topic really resonates well with, you know, what I've faced over the, you know, more than a decade and and how we are still trying to work on with all despite all the good effort that we've put in place.

So really excited to be here.

Good morning, everybody. I'm Mike McMillan. I'm the manager of asset performance at Nova Scotia Power. We're up in Canada, on the other side of the the continent, the East Coast.

Part of my role, I'm within our enterprise asset management team.

So we're responsible for, you know, setting asset strategy and as as my boss likes to say, the business connected. So important really important topic today to talk about silos and how we can share the data that we do have and and make better decisions.

Thank you. Morning, everyone. My name is Craig Stenberg. I'm with Southern California Edison. I'm a I founded the remote sensing and unmanned aircraft department center company within our aircraft operations.

And I've been with the company for about fifteen years now. And what we were able to do over that time is bring our remote sensing into the twenty first century. So happy to be here and answering any of your questions.

Great. Thanks, gentlemen. So as our keynote speaker stated earlier, I was really excited to see or to hear some of the stories that he was telling because I think early detection and prevention is really in today's day and age to catch incidents before they happen is really critical. And I think no matter what department you sit in, it's to to quote him, outages are bad. Speed is critical.

And I think this is something that's gonna be really exciting as we dive in. So for all three of you, and I know you've jumped on this question, so utilities generate enormous volumes of data. So you have spatial, operational, asset data, but much of that value remains siloed. So from your perspective, what's at stake if these silos persist, and what becomes possible if data is shared across functions?

I can start there. I think the the biggest risk is if the data isn't shared is you make bad decisions, and bad decisions could mean you're investing in the wrong assets, could mean, in the case of emergency management, you're restoring not the wrong customers, but maybe customers, you're missing that critical infrastructure.

If you if you can't get that data sharing right, if you can't break down those silos, you're gonna be wasting a lot of time and a lot of customers' rates and, you know, potentially putting you know, increasing your reputational risk. So so that's kind of the ones the ones that I would see.

Yeah. I was going to get the same answer, and and especially, like, the ultimate thing is it impacts the customers. It doesn't matter what happens. Right?

You know, whether it's an asset issue or a wildfire, you know, or or any kind of an outage, ultimately, it just impacts the end user, which is the customer. But there are these data silos also what it happens is that it will ultimately create a trust issue with the different programs that are being launched or new programs because you you see the data coming in, it's very siloed. Others don't understand where that exists or how to leverage that data. So as you try to create new programs, like, let's say your company doesn't have a drone program today or any other like, a geospatial program, and if you're trying to prove it, the past experiences experiences will harm your ability to actually scale these programs.

Right? So the data silos are are are a big problem. But, again, even if you solve the data silos, the next big challenge would be is how do you leverage all that massive data into insights as well? But to get to the insights, you need to be able to make sure that these datas are communicating or in one location, and then you can move on to the next thing.

And I think one of the challenges we've had at Edison with this is that when you're not sharing the data, your other departments go out and capture the same data Yeah. And at a usually great cost.

We have six helicopters that serve fifty thousand square miles of territory. So when we get a request to go and capture some data, and then three weeks later, we get another request to capture the same data, we don't really have a lot of resources to spread around to do that. So so data silos has been an issue.

It's an issue that we're now beginning to address and fix, but I think it's gonna be a little bit of time.

And as I understand, incurring helicopters are kind of expensive. Right? So if you had to send it twice for the same inspections?

Yes. They are. I would say, on average, we'd probably pay about four thousand dollars per hour to operate a helicopter.

So So definitely not worth doing twice when you could do it once.

I'm really excited to be up here with these three because they represent three very, very different functions and very different departments who often have to work together. But historically, I think it's just been very hard. Even in my experience, I started my own career in generation, and after thirteen years, moved into asset management where I was like, oh, there's this whole thing called transmission that I didn't pay attention to for over a decade and kinda had to learn. So I have a few questions for each of you in the functional for your functional perspective. So first question is for Mike. If we talk about asset performance, from an asset performance standpoint, what data has the biggest impact on reliability and risk today?

Yeah. Great question. So I I think the biggest one for us is anything that we consider a leading indicator. So not necessarily I I mean, I love analyzing average data as much as the next person, but, you know, you're already behind the eight ball once you're looking at that data.

Historically, that's primarily been where we've made decisions, then you you end up chasing problems after they've they've occurred, and you can't really be proactive. So from an asset performance and asset risk, may mean better risk based decisions. It's all about being indicators. So things like our inspection records, things like condition assessments for vegetation, all of those that we can really leverage that data.

You can be more proactive. You can prove to your customers, you can prove to internal stakeholders, your regulator that, okay. This decision that we're making is based on that good data. It's before we have a problem, hopefully, and we can we can kinda prove that we're making the best decisions in the interest of those in those customers.

Great. And you are kind of the end recipient of getting data. Right? I think we're in a place where I think everyone has accepted that more data is a good thing. We don't have enough of it. But once it's collected next question is for AP in strategy planning and governance. So how do strategy planning and governance shape how new data sources like UAS, drug data, are adopted and scaled?

I love that question. So, yes, I mean, strategy like, if we if you think about long term, right, like, when you when you think about the new data sources, the strategy should come from what does the business need. Like, yes, we have a fancy tool or a very effective tool. And if you don't know what you're gonna use for actually or why you're gonna use that, it just becomes an expense operation.

Right? And so you need to be able to latch it with what is the business outcome and expect it. What is the business problem that I have today that I'm unable to solve? Or maybe sometimes it's all about what do I don't know today?

Right? Sometimes we don't know stuff, it happens, and then we are spending hours and hours investigating after the fact, okay. Why did this happen? We thought we had this.

Right? Sometimes it's lack of information. So defining that business need is very critical, and that's where the strategy comes from. And once you have that established, it also helps you in standardizing, okay.

What do I need to capture? Then comes the planning side of things where you say that, okay. Now that I understand, I need something from this particular source. How often do I need it?

How differently do I need to need it? And how does it change from what I'm already gathering today, or how is it different than what I don't know today?

And lastly, what happens is comes down to once you decide on all these things, is the governance space, which is the most critical, I would say, is because now you need to have a data owner. And how does this data sync with other systems? And as it's syncing, how is it changing your decision making capability as well? So all of these things, I mean, you can it's on paper, it seems easy that you can define everything ahead of time, and sometimes we have to to support these programs, launch these new programs, and scale them.

But as you go with it, you need to be agile in also tweaking that to say that, okay. This is the initial assumption we started with. And as we get this data, actually, it's way more reliable than what we expected. Or or maybe we need to change some things to capture more enhanced data.

Because if you don't have such a plan, your drone operations or your drone photography just becomes a very expensive photograph that you're taking of the same assets that you could have just done from the ground. So you need to be able to differentiate between what do we have today and how is it gonna change that outcome. So and going through that experience myself for three years, launching the first growth program in all of Exelon. Right?

Technology one thing I learned is that technology will enable adoption. But if you don't have the right strategy or planning or governance, you cannot scale the program. And that's where we focused a lot in the initial stages. Like, even six months to eight months, we were only doing very minimal missions.

We were establishing the standard operating procedures, the emergency procedures, and things like that. And everyone was like, hey, we funded you so much. I haven't seen anything flying around that much. What's happening?

And we're like, you gotta wait for it. And once we started that, it completely changed the perspective. And we have about thirty, forty plus pilots, million or a couple of million dollar plus savings, and we are doing drone operations, like, every day. We pretty much replaced ninety percent of our big offer inspections as well as the drones.

So it just comes from how do you build that strategy and how do you plan and scale it.

That's really good because I feel like technology, any sort of innovation, kind of starts with a science project. Right? I think utilities are in a space now where we're kind of past the r and d phase. And admittedly, the technology also needed the past five, ten years to come along. Realistically, the drones that were available ten years ago are not the same as what drones that are available today.

And so I really like what you said because I think the goal of scaling and operationalizing once you get past just doing a POC just becomes so much more complex. Absolutely.

And if I may add one more point, the the key is is start small. Like, you know, when when we looked at the drone programs that the other companies were doing or even in Exelon, like, the other operating companies were doing, and I used to be in Baltimore Gas and Electric, by the way, when we launched the drone program. So we have other operating companies with an Exelon.

Everybody started with the most ambitious and the difficult use case, which is the storm management. They wanna use drones for storms. And when we went back and as we're strategizing, we said that, why did they fail? I mean, they've been doing this for two years.

Or why couldn't they really scale the program? And it became like, hey, you need a lot of people to do that. And storms happen only twice or maybe once a year. Still bad, but it happens that.

So you don't get to test your program. You don't get to build that trust. So what we did is different. We started with the most easiest programs.

We build a trust, and then once we build a trust, we started recruiting more people into it. And within a couple of years, what happened is we had a pretty much an army of drone pilots. So when the storm happened, they could go the entire system and scan the system way more faster than you could have ever imagined.

So Yeah.

And this is something that we're seeing across the industry. I think that reactive maintenance decisions using drones is what is in the news. Right? And our goal is to stay out of the news.

So when you think of storms and fires, those are the two main things that nobody wants to be in the news for. But how often do they happen as opposed to everyday corrosion? Everyday bolts falling out, vibrations, just stress corrosion, cracking, fatigue, woodpecker holes, all of those things. And those are the things that if you can catch those early indicators, it would make a much bigger difference in your proactive maintenance strategies, which ultimately you never if you're doing it right, you never get in the news.

Right? But I think it's hard to see that when the sensational big storms happen and and you have those responses.

No news is good news.

Yeah. So, Craig, this next one's for you. So from a UAS perspective, what does it take in to elevate a drone program from a tactical tool to an enterprise data source?

Well, kind of to I didn't know how pronounce the name.

N. AP. Okay.

To AP's point, you know, starting small.

And in the early days when somebody came to us and said, should we start your own program?

That was something I think it was our CEO at the time, which means start your own program.

That was you know, we we started with one drone and and one guy flying it. That was me.

And, you know, which is really how do you learning what we don't know.

You had to know what you don't know, at this point.

And, eventually, you know, we had our first, you know, dozen pilots or so, and we realized pretty quickly that if we're really going to scale this at all, we're we can't do this as a centralized operation. And so that represent another problem that really it's not technology or, I guess, the the scalability of the aircraft itself. It's really more about the operator. So we're an aircraft operation. We like I we we have helicopters. And every time we bring in a helicopter pilot that we're gonna hire in, that this person has at least three thousand, four thousand hours of flight time, years of experience, a tremendous amount of knowledge in aviation.

And what we're looking to do now is take somebody with the same kind of skill set on the alignment side or the, you know, the utility side and introduce an entirely new concept on how to operate. And, basically, we're gonna try to make professional pilots out of people who've never flown anything before. So that was probably the biggest challenge that we had as far as scaling. But and and and also the we we mentioned that the drones today are much different than they were ten years ago when we started all of this. So it's you know, and and thank you. I I like to say optical has been a huge help, but it's not a flight plan, just so you know.

But it that was being able to bring in equipment that would be much more reliable and also had an it's an ability to avoid the objects that you're trying to look at was all was another big piece of this.

What we did then is by crowdsourcing the aviation, we've actually gone from a few dozen about a dozen pilots initially to about a hundred and twenty today. So we've got about ten x increase. We actually even have reduced our size, believe it or not. We we're about a year and a half ago, we were about two hundred pilots in the company, almost two fifty.

And when we transitioned to our current platform, we were significantly higher cost platform. And as we're bringing in the equipment of the the old pieces that were now six, seven years old, some of these were brand new in the box. And we discovered there are lot of people who joined the program because they want to be a drone pilot, but they actually have no use for the drone. So so that's that now we've gone through kind of that roller coaster.

We've now kinda I think we've rightsized to what you know, our and we're a large utility, so our rightsize might not be your right size. But we've right sized that program, and the other piece is that we have different organizations in the in the company that have different mission sets. So by having a centralized governance on how to fly the aircraft, which is very basic, we now allow other organizations to as long as you're staying within those rules, you write your own rules on what you need to do. Yes.

And that is the other piece that really helps synergize the operation. So which synergizes the data too, and, hopefully, that'll kind of set up queue the next phase of this.

Yeah. And I think what I'm really impressed with with Southern California Edison and how you guys have set up your program is that you actually give a drone to each pilot. Like, your pilots are you didn't incur a lot of headcount hiring hundreds of extra pilots. You made pilots out of the people who need to see assets anyway, or at least that's what I'm hearing. And I think the work that we've done together, just seeing how somebody in generation has very different inspections than somebody in TMD has very different inspections that they use the same drone for, but you've in the aerial program, really train them up to be able to use the drones as a tool. So I've I found that really exciting at your program.

Oh, thank you. That was you know, that that did take some doing to get there. We had literally had to build the training program for it. But the other half of that that you're very modest about was being able to partner with a technology a technology partner who will respond to the needs of the company and develop the basically develop the next work methods on how we're going to do this, which is a huge technological advancement in order to do that.

It's and it is important. If you're a vendor working with the utility, it's important to be flexible and be able to move in the direction utilities may need to move because and maybe as we're fine throughout the as we move forward with this conversation and additional conversations here, is the more data that you bring in, the more insight you get from that data because it it synergizes to other pieces of data. And that was part of what helped develop our program is we're doing this. We're like, hey. What if got a drone that can do this? And having a partner who's willing to build that out has has been pretty major for us.

Thanks, Greg.

Partnership is really what we're all about, and it's been great because I think, like AP said before, your solution is only as good as understanding the problem you're trying to solve, and I think that's what we're all trying to do.

So my next question is for everyone. So what barriers most commonly prevent data sharing across teams, and what has actually worked to overcome them?

What?

I think for me, the biggest barrier has been, like, knowing that data exists that that people could use. And we've seen lots of examples, you know, and Craig mentioned this too. So, you know, we collected this data and then somebody else collects the same data, and maybe a third person collects that data.

So so having the ability to kinda sit at the middle or to understand, okay. This is what other folks are doing from a remote sensing point of view or just, you know, traditional inspections. What data do we have out there knowing even who to who to talk to? So one thing that has really worked for us is kinda building cross functional teams that are meeting on a regular basis.

So, you know, distribution, we've got so many stakeholders from folks in the field that are doing construction and maintenance to, you know, regional engineering to asset management to and on and so on. But even just to get some of those people in the same room to say, here's what I'm working on. Here's the problems that I'm having. Oh, I might have that data for for you.

Oh, we collected this loading data or this imagery that might be a good fit.

Just sparkling those conversations have been really important for us to understand what's already out there and what we can leverage right right off the hop.

So as far as data sharing goes, I think what we've found in, my experience anyway was, one, everybody wants to control the data, but nobody wants to own it.

Oh, really? Yes. That's nice.

And and the only thing I can really put if I were to try to pinpoint what's going on with that was it really had more to do with ego, I think, than anything else. Everybody wants to be the one that pushed the button, but but, you know, no and they don't want anybody else to be able to push it. And so as a result, that's what I think prevents people from wanting to share their data.

But at the same time, all the questions people are asking require all this data that is there. I think it's what's driving all the you know, well, you won't give it to me. I'll go get it myself. So we're finally starting to break away from that at Edison, but it's actually taking quite a bit of doing.

And it's in an odd way, the data that you're collecting, it it almost mirrors socialism in a way. If everybody were to collect the data and then put it in one repository for everybody to use I'm not a socialist, by the way, but it's just a model still. It's there. It it puts that same repository there for use, and everybody can use it the way they need it.

And and, you know, instead don't know why that there's been so much resistance to that, but the point where our information government's living company is starting to say, I guess, we're gonna have to be the ones that own it.

So I think if I double click to give this question to AP, how do you build trust and share data and decision when decisions have real operational and regulatory impact?

Looks like you read my mind and just asked the question.

No. Yeah. I was gonna touch that. And I agree with both your points there.

That's one one challenge we've seen is, like, you know, sometimes data makes people scared. You know? They are suddenly seeing so much data that they're not privy to in the past, and they don't know how to react to that data. Like, I'll give an example. We we we have a very robust typical pole inspection program. And as we were doing a parallel inspection to to a point with the drones, when we captured those images and sat down with the typical inspector and went through those images, our work orders shot up by four hundred percent.

That doesn't mean that the inspectors who were doing it were doing it wrong in the past. It's just that you didn't have that kind of a lens until at that point. And suddenly when that came up, our work orders shot up and some of the critical even within the work orders, we have priority tens, twenties, thirties, different priorities. When you have a highest priority, you need to fix it within a week, basically.

It's a reportable thing to the regulators. Now suddenly, we had to pause the entire program saying that we're not equipped for this. So if you know something, you gotta go fix it. And then the question is, what is the right thing to do?

Now we are realizing that we may not have been getting all the data that we want, and suddenly there is this device, but it is not inducted into the full end to end preventive maintenance program either. So we had to really take a step and rethink our strategy completely. And then after that, what happened is the transparency piece. Are we able to get the same data consistently?

Who's actually capturing this data from the drones? And and and how often is it gonna happen. Right? So we have to define the transparency to say that this is how exactly we capture this data.

This is the time period we're capture this data. So it it creates an auditable trail. So in case you find it very critically short, you have to take down a service or something happens, unfortunately, you can go back and say that, yes, we have identified at this point of time by this qualified person, and here's where it's stored, and and we can actually work with the regulators on those aspects. If you have to take down a line on a temporary basis because you just recognize the highest priority issue happening on a pole that might just take on a transformer or something like that.

And then comes, again, the governance piece. Like, where is the data going from there? Like, you you've captured it. You saw the data, but where is it go and how people are gonna leverage that thing.

So that's another big silo that we had to break down as well, which was very time consuming, but it was worthwhile to go through that effort.

In Exelon, we have launched a data platform or big big data or data lake, whatever terminology you wanna call that, in twenty sixteen. So that changed the game for us when it comes to data. But one thing what we realized is that we are capturing millions of data points in a given day, but now how do we extract insights? The silos are broken, but the question was, did it really break the silos or it's just that everything is being dumped into a storeroom now and you have to go to the storeroom to really find out what that data means.

Right? So and then now we're adding this drone data, which is overlapping with your existing data. So what how do which data holds the truth? So it became a a a planning to trust the data.

So we had to first create how we're capturing it, and then what is the transparency of the data, like, which means, again, as I said, who's capturing it? What what is the different process to capture that? And once it happened, we had to do a lot of field validation to say that I saw this split on the crossbow, which means it's gonna come down. It's a priority twenty for us.

And then a week later, the fleet crews would go in and say that, yep. Yep. The image was right. It's very accurate.

It makes sense to us. And we have to go through multiple iterations of that to build that trust. As it built, we were able to secure the data and put the right governance, and then everybody started using that data as a source of truth, and it started ending up in our system of records. So it's it's a journey.

You won't get there day one, but you gotta plan for it. And you gotta figure out what is that gap, what is that data process that you have within your own companies, and then realize, okay, are we do we have a data lake like what we have, or is it siloed? If so, who's your biggest stakeholder, and what to do with that?

If I could follow on that, who are the key stakeholders that you found that you had to bring to the table, right, when you had all this information and suddenly there there are decisions to be made.

Once we have the data getting to the data, flying the drone is the easiest part, I would say. But once you land the drone, the difficulty begins actually. So it was mainly the executives because suddenly, they they did not plan for this. They did not plan to have this information.

As I said, suddenly the work order shot up and you don't have that much workforce available to get this done. And even if you can go contract the work source, you don't have that budget because you didn't plan. You're planning to see maybe a thousand problems in a given year, and you plan for twelve hundred, and within the first quarter of the year, I already showed you three thousand problems, then you're not planning for it. Right?

Convincing them, how do we standardize this? And also working with our teams that interact with the regulators to say that, okay, this is a new form of doing inspections and how do we really change our protocols around it, our preventative maintenance protocols around it, and things like that. And then ultimately showing why is it worth the investment. Now that you have identified so many problems, we were looking at retrospectively and saying that, okay, is our definitions even true?

I mean, if every poll is going to have ten problems, then is it really a problem? Because they did not come down in the past twenty years. They're they're up and running. So what does it really mean then?

So we have to go back and define how do we take these data points and actually draw a more specific derived insight to make that judgment call on whether we need to go do something about this or not. So those were the biggest, you know, folks I would say we had to convince. And then at the same time, it was also the field operators because our when we launched the program, one philosophy we started with is that we are not going to give drones to anybody just like that in the company. We are only going to give it to the folks who are actually doing that function.

Like, you know, if it was alignment, we're supposed to get on a bucket truck and look at an equipment, only they're gonna get a drone as a part of the enhancements, like, you know, or or maybe a transmission operator. So an engineer wouldn't get a drone because a typically engineer will not be going into the field to inspect something on their own. Right? So we created that pipeline of things and we had to ensure the field workers were trusting the technology in their hand to make to ensure that they can get the information what they need.

So it's it's it's again, it's it's a stage of those. That's why it took us couple of years to get to that point.

But mainly convincing the field folks that this is the right data source, and once you have that, convince the leadership on how to act on it.

Awesome.

I have a question for Craig. So how do you maintain consistency and accessibility as UIS programs scale?

Well, I kinda answered him a little bit earlier, but I'll go I'll expound on it. So our central, I guess, operation for the drones within aircraft operations would be well, we're using what we call a model manager concept. So if we're the sheriffs in town, we'll put a few deputies around the company, and they become kind of the chief pilots of their organization.

And and it's it's that model manager thing, if you go right across the bay here in North Island, that's where I kind of developed my operational knowledge is I was in the helicopter squad over there.

You remember a game back in nineties, a computer game called Wing Commander, that's a real thing. There actually is something called a Wing Commander. And if you're on the base over there, there's a helicopter wing. That's just one little building that actually has no helicopters attached to it.

But if all the people in there are managing the programs, everybody else is flying on different squadrons. So the, the wing commander would be the commanding officer of all the, squadron commanders. Okay? So we just took that cookie cutter object and put it right into Edison, and that's how we're operating that program.

So what that's done for us is, by giving the basic skills of flying the aircraft to the people that are flying them and then developing the the higher technologies and developing standards that they're going to have to meet in order to do these more advanced operations.

We've empowered everyone else to, find the best ways to use that new tool.

And that instead of trying to be, you know, dictatorial, you know, this is how you're gonna use it, this is how you're gonna use it, it's it's really not about what you're allowed to do. It's what more about what you're not allowed to do. Do whatever you need to do. Just don't do these things. And that's really unlocked the the creativity within the company. And we're we knew going into this that if we do that, we're people are gonna think of ways to use these things we've never thought about. We just have to plan the rules in such a way that they don't come up with new ways to screw it up that we've thought about.

So Gargoyles are definitely important.

So I think we're getting to our last ten minutes. Todd, I am watching the time. Don't worry. So for you three gentlemen, how do you measure success for cross functional data initiatives? I think AP touched on the need for executive support and alignment to really bring some of these things to scale and to align across departments. But, really, what what metrics do you use to measure success?

I think one the things that we can look at if we go back to the early days of our program was one way of looking at it was a successful week was a week where nobody smashed one of their aircraft. Okay? That was which was very few and far between. And we get to this end of it where, you know, we've a lot more developments happened, but we've really improved our internal training. The equipment has improved significantly.

And now it's really, it's more like, hey, we get through a year without anything happening, which is you know? So and that's the operational way of measuring it. And the other, I think, probably more significant is, you know, a lot of our use cases are centered around wildfire prevention. And it's like, well, you kinda have to go back historically and say, you know, back at this time, all these fires were going on, and you fast forward to today.

And Craig Fugate kinda mentioned some of this as it it's also the early detection piece. But on the prevention side, both of we've seen both of those. It's not just been the drone too. We've we've I've done other infrastructure improvements to help prevent wildfire.

So now it's gotten to the point where we have very few wildfires, and they do occur. And it's a more difficult way of measuring because it's you're it's where you're measuring what how what did you not lose? You know? That's not, like, the same kind of metric, but it is one way of saying that because we have we're we're diverting so much fewer resources to recovery, they were our prevention.

That has been, I think, one way of measuring that.

Any more?

I think I think for for us, like, a lot the great great points there, folks aren't smashing what you built.

That's that's usually a good a good sign. One thing that we've we've seen as success and maybe a a difficult metric to measure, but, you know, we are building tools to analyze risks. For example, in the transmission system, we're bringing together, you know, our inspection results, our vegetation results, our lagging indicators like performance, reliability.

And when we start to hear that people are teams that, you know, they don't report to me, but we work closely with them are saying, yeah. We we use the risk dashboard, we identify these risks, and we build a project, and the project's in the capital plan. Like, I'm I'm crying from joy to hear that. That's that means we were successful.

We built the trust. We built the right data pipelines. We brought the information together in a way that can be consumed by those teams. And when operations and capital and folks who maybe are only seeing, you know, they might have the temptation to just focus on, well, I saw this pole was broken, and I'm gonna go fix but I don't have the the view of the full scope.

When they're starting to use those tools, that's a huge win for us because it means we're making better decisions. We're not being as reactive. And when we go to talk to our customers and other stakeholders, we can stand behind those decisions and say, yep. This is why this is priority number one for us.

We can go to our regulator and say, this is a capital plan we stand behind. You can be confident if you approve our expenditures. So so that's a huge that's a huge win win for us. And I'll I'll I'll give you another example.

Like, we have operations teams and folks, you know, not disparaging anybody working in operations or working with field folks, but that is a hard nut to crack to get them to use maybe a a different way of working or or something new that they've seen. And when we talk to them, and it takes time to build that trust, we've talked about that, But giving them tools that really resonate with them, and and again, when you get a call from somebody that says, hey. I'm looking at this risk score. I'm looking at these inspection results.

Can we dive deeper into that, and and we can talk about why this is red. This is your number one priority. Well, you you've saved a whole lot of time and heartache. You've saved crews from making unnecessary trips.

You've potentially avoided outages. So so that's another huge way that we're we're looking at measuring success.

Very nice. We should talk about that. Yeah. Okay. There's a few folks. We I mean, the program's a bit more mature, so I'll start from, like, how we started off, you know.

So we created a rubrics to just to convince executive leadership to say that excuse me. How is it made it comes down to the safeties? That's the language that everybody understands very easily. Right?

So we talked about how much cheaper it is going to be using drones compared to the typical traditional inspections. And then we said, what about the worker safety? Like, you know, how will it improve the safety of the lineman or the field person that's doing this thing? And then we talk about our operational efficiency.

Can you do this more often?

Excuse me. Sorry about that.

Operational efficiency. Can you do this more often? Like, you know and next comes the data quality. Like, you know, is it a better resolution data?

Will it help you make better better analysis or or better outcomes. And then once you meet that, each individual use case will have its own detailed benefits, cost benefit as well. So I'll give an example. We had an issue about where we had a lot of water crossings.

The lines were doing the water crossings, and we had to install these tags on the line so that the boaters or the sailboats won't hit those lines. When we did with the traditional methodology, that was costing us almost like one point five million dollars. It takes five years to do that because we have to put a bucket truck on a barge, and the barge has to go into the center of that crossing, and they have to go up with the water shaking and the wind and everything. And we said, why can't we do this with the drones?

So safety was paramount. So safety conditions went from one safety to five safety because there's nobody in the water. Somebody's on the bridge or on the side by the bay, and they just climb the drone. And from a savings perspective, instead of spending one point five million and accounting for losing two or three drones into the water because of any kind some condition that we cannot control, all of that together, we spent a hundred and twenty five thousand dollars.

That's it. So we saved ninety five percent of the cost right there. So we did that analysis for every single thing. Another big example is that we have these natural gas towers or silos actually in the downtown Baltimore area, which has to be inspected every other day.

Somebody has to climb up all the way, which takes them three to four hours, and once they're up there, if there's a high wind leaving, they gotta sit there. They cannot come down. And it's a it's a huge physical fatigue. It's a safety risk, and it's almost one dedicated person just doing this.

And we were able to do the same thing with the drone in ten minutes, and instead of every other day, we can do it three times a day now. And what we also did, thanks to the drones, we actually automated the whole damn thing. Now we don't nobody has to even go out into the field. They just push a button from the control room.

The drone comes out of the box. It does everything and comes back, and the data is being processed by AI in the back end. So so what what happened is, like, or what we did is each you need to establish we established the baseline saying that these are the criteria that we're gonna meet before even we launch into a use case or go do this thing. And once we do that, each individual use case is different.

And another aspect what happened is people who were rolling out with, you know, the mid sized bucket trucks, they don't do bucket trucks anymore. They just use a pickup truck, actually, with a drone or a couple of drones, different drones depending on what their function dictates. They just fly the drone. So instead of purchasing a mid sized bucket truck, which is few tens of thousands of dollars, we're buying a simpler pickup truck with a drone, and we are able to do that.

And with our sustainability goals, what we did is that we gave all these field folks EVs, EV trucks, Rivians. And we said that this is a completely zero carbon inspection. So we changed the narrative to say that this is a very environmental friendly inspection as well, and we can do it as many as we want. Because drone is powered by batteries, EVs are powered by batteries, and you change the narration.

And, again, the PR concept with that too.

That's awesome. And I know we're running out of time, so I wanna wrap it up. But really wanna thank you, gentlemen. I think the future is really, really exciting, and I think what's more exciting is the future's here now.

So I think AP touched on not only the value of flying drones, but the idea of flying drones remotely. And this is something that Skydio is really, really passionate passionate about. So if you have time in the next few days, we're over at booth five four five zero one. You can fly a drone yourself remotely in different parts of the country, and we also have this Paraverse arcade simulator that's specific to utilities so you can conduct some utility inspections and see how good you are as a pilot.

But overall, we would love to talk to you. We're very excited about the future for utilities and just being able to automate and make everything safer, more reliable, and more resilient. Thank you.