# Building a Utility Drone Program for Critical Water Infrastructure with Denver Water - Event: NestGen ’26 - Published: October 2, 2026 - Length: 34:52 - Watch: https://nestgen.org/library/watch/MlXWsPo_1d0/ > The transcript of a recorded NestGen session, for reading or for use with an AI tool. It comes from the video's > captions; automatic captions can mishear names and terms. ## About this session Water utilities manage sprawling, critical infrastructure that's costly and time-consuming to inspect on foot, and any disruption carries public health stakes. Utility teams need a way to monitor reservoirs and distribution infrastructure faster, without pulling crews off other priorities. In this NestGen session, Denver Water will share how they built a drone program for critical water infrastructure, and what utility teams should think through when scaling autonomous inspection across essential systems. ## Transcript **[0:01]** All right. Hello everyone and welcome to another session at NestGen 2026. I am Neil and I'm the head of customer success at FlytBase. And today during this session uh we going to do something different. So rather than the traditional presentation approach that we've been doing today, we're going to have more of a conversation about what it actually takes to build a drone program inside a water utility. a place where the infrastructure is critical, spread across large areas and the technology has to work as part of the operation. And with that, I'd like to introduce our guest Brian Frasier as well, who leads the Demo Waters utility uh pro drone program for all of their critical water infrastructure. Hi Brian, welcome to >> Hello Neil. >> Hello Neil, thanks for having me on. really glad that you could join us. Uh I'll just give the audience a little context **[1:01]** about you and what Denmo Water has been doing and then we can start with our session. >> Sounds good. >> All right. So Brian has spent the past several years building and scaling operations that span, you know, the entire assets and reservoirs, treatment plants, pipelines, and distributed networks. uh turning drones from let's say one-off experiments into a core tool of how uh water utility monitors and protects his system. Uh before to taking on this role, Brian worked across survey, inspection and field operations and uh now that gives him a rare combination of hands-on experience with assets themselves and a clear view of how technology can make those inspections safer, faster, and more repeatable. He's been at the center of questions like where do we start the drone program from, what to fly first, how to navigate security and regulatory constraints, and how to make a compelling case to the leadership when you're dealing with critical **[2:05]** infrastructure. And uh separately on a personal note very excited to talk with Brian because uh Demo Order sits at the forefront of how traditional utilities are adopting autonomous systems not again for just cool demos but to solve real challenges around safety, compliance and long-term asset health. So once again, Brian, welcome to I'm really glad to have you here and you know I'm glad that we we got the opportunity to host you today. >> Yeah, it's pleasure to be here. Um I'd like to thank the audience for joining in and uh look forward to hopefully sharing some knowledge that I've learned uh over the past it's been we're coming up on two years um trying to get this program going. So um yeah, look forward to share. All right. So before [clears throat] we begin, just a couple of housekeeping rules uh for the audience as well. So the session is going to be recorded. Uh if you have any questions while Brian and I are having a conversation, feel free to drop them in **[3:10]** the Q&A section and we'll pick them up at the last 5 to 10 minutes of the session. If you like what we're talking about and if you feel this resonates with you and probably somebody else in your network, take a screenshot while we're presenting, share it on your LinkedIn and you know use the hashtag neesgen and um you know hope you learn a lot today. It's going to be an insightful session. Uh Brian, I'll let you take it from here. Uh I know that you've prepared something from your end as well. I'll stop. >> Yeah. Yeah. Yeah. I will uh I'll jump we'll we can jump into the questions and I'll I'll share I think um I've got some kind of de some examples some some [clears throat] graphics and stuff that I'll kind of highlight um >> as we kind of run through these questions. So >> um but yeah I think maybe we just start off. >> Perfect. So I think let's start with the first question and you know this is something that I personally do uh you know start with the why. So let's let's pick that one up. what made Denmo Water look **[4:14]** at drones in the first place and you know what was happening on the ground that actually made you think that there might be a better way to do this. >> Yeah. So let's just a little background on you know Denver Waters assets. Um we like you again we we span a really large geographical range. Uh we manage 17 reservoirs, four treatment plants, 21 pump stations, and hundreds of miles of canals and pipelines and diversion structures. Um all these assets combined to supply water to roughly 1.5 million customers across the Denver metro area. Uh historically, a lot of our sites, our facilities have been managed by uh man operations. Uh we have 247 coverage. Uh so a lot of these res remote live on site year round at these um so that requires anywhere from two to three up **[5:16]** to four personnel on site. Um and these people are in charge of everything you know maintenance you know fixing electricians plumbers they pretty much have to run the gamut of all trades um and have the technical knowledge of our water system. So they're really uh really trained, highly specific individuals. Um uh so kind of what led to this conversation is we were you know we were losing uh people. People were retiring. A lot of these facility managers had been working up there 20 plus 30 years. Um and as those facility managers retiring it was hard to kind of backfill um and find employees. Um so the organ organization had already kind of been moving towards centralized operations from our admin. Um you know with stuff like SCADA controlled valves which is a remote operation systems to open and close water valves. Um so this kind of beyond visual line of sight drone **[6:19]** um pilot program kind of fit naturally with kind of the progression of the the organization. Um so we decided you know hey let's why don't we give this a try at Antaro um which is this um image up here. This is Lake Antaro. Um and you can see it's it's way up in the mountains. It's super remote. Um a large property also to manage as well for just one you know a couple individuals. Um, so pitching this for uh to have a drone, a remote drone up here operating was uh was pretty easy as far as like the the vision of it. Um, >> I think those were all my Yep. So the just the sheer you know expanse of the area that you have um you know that was probably you know an easy cell or like a no-brainer in a way that okay we need drones to go out and inspect or surveil this particular area for all of your assets. **[7:24]** >> Yep. >> Yeah. So that's Yeah. So that's the kind of um so the value that we kind of were pitching um was that these were you know this is not this is a multi-tool functionality. It's not just hey this is for inspection or to supplement um you know just eyes in the sky but you know with me having being a surveyor and having a survey background I had already been using drones for mapping um purposes. So I kind of saw that side of it and I worked closely with our director of security and our director of SOS who manages all this uh our water collection system. Um, and so they were saying they were coming at it from different angles from the security standpoint, from the inspection and asset maintenance standpoint, and for me, my survey background, and also being a I had the regulatory experience cuz I had already gotten my license back in 2017. **[8:24]** Um, so I kind of did the heavy lifting on the regulatory side, too. Um, so we kind of had kind of the three main pillars of this program, kind of heads of it. um kind of putting our minds together to to kind of get this program off and going. And that's kind of how we shared it um to the our chiefs was that hey this is you know this tool can be it has multi-functionality multi-utility um with one system you're you're gaining a lot of lot of different utility >> right and you know you mentioned security inspections and probably surveying as well so where did you actually start like what was the first let's say use case or first problem that you figured out, okay, hey, let's let's tackle this with drones. >> Yeah, that um so for me it was the Antaro opportunity came up pretty quickly um because we actually didn't have we had one caretaker there and they ended up **[9:28]** leaving and we didn't have anybody up there. Um, now the site's not it's not a high security site per se, but we do have some we have some issues with because it's an access to public. Um, and mostly, you know, a lot of fishermen um utilize that lake, but they we do have a restricted access area. Um, and it's more for public safety, but we don't want people fishing around the the intake. So that antero opportunity kind of thrusted this program into into you know the eyes of the chiefs. They're like hey we got to we got to take care of this. So that kind of gave it supercharged the project. >> Um so the antero was really the like kind of spearheading the need. Um but there was also a lot of other opportunities that I saw with survey. This one you know the image up here this is our cheeseman dam. Um, you know, regul regulations have been getting tighter from the state to inspect these structures. Um, we are currently required twice a year to report uh to the state **[10:34]** [snorts] and you know before like it's it's it's really cumbersome to do these inspections. Um so drones kind of naturally fit um in that space as well as as an inspection tools for you know you know really large infrastructure and hard access infrastructure. Um so yeah those those were kind of the two that were in the forefront which we had already kind [clears throat] of identified early in the process. So I know we've had a discussion regarding this but uh just so that can understand exactly what the challenge was. Can you help us you know maybe paint a picture on uh how this was done before drones and versus like when you actually had drones to inspect this kind of an infrastructure. >> Yeah. So the the challenge um so previously what we were doing with uh dams is we were actually repelling. Um we were using ropes and harnesses and repelling down **[11:37]** these faces to do visual inspections. Um inspectors would take notes, detailed notes, um take photographs and then they would create a you know a detailed report and give that to the engineer to review. You know this process would take months. Um, and it's still you're talking really only a handful of data points. Um, where with a drone you can get thousands, hundreds of thousands of data points um, and do it within a in a day. You can complete the inspection. Although, I mean, you still need, you know, visually you take the model um, and visually inspect it back in the office, >> right? and have my this video. This is kind of a little video that highlights the what we were looking at using um essentially the platform, you know, it it works really similar. uh you have the **[12:39]** photoggramometric model and then as you're clicking and and looking around um the feature then it identifies the images that are tied to that photoggramometric model. Um so the drone combined with tools like this um it kind of really gave a visual of okay the engineers are like okay I can you know this seems a lot more feasible to do a complete detailed inspection um >> um on the dam. So >> but yeah that is a that's a great uh great leading question there Neil. Um so yeah the challenges that we still you know also like canals all these pieces of infrastructure are extremely large and there's lots of concrete and lots of just different aspects to it. So to really get a full view, the drone has really kind of unlocked that ability uh for us. And now a lot of our engineers **[13:41]** are really like okay this is seeming less daunting to try and manage and you know report to the state um on the condition of these assets. Um so >> perfect. So I think this is where you started seeing the value of drones and you know autonomous drones is started giving giving you those uh quick and early wins and uh >> did this also help you build the case internally for maybe expanding the drone program uh particularly around like other use cases or maybe some other internal factors like safety cost savings etc. >> Yeah. Yeah. Obviously, a big, you know, safety is is at the forefront of a lot of the work that we do. Um, you know, a lot of our technicians and employees um have to be in in really um kind of tricky terrain, you know, steep mountainous terrain. Um we also have a lot **[14:43]** of enclosed spaces. Um so the drone I think really started now a lot of the newer drones have um ability to fly without GNSS. Um so we've been kind of looking at those systems to do um confined space entries, tunnels and canals or tunnels. Um we have a lot of underground vaults that house a lot of the valves structures. Um and in our standard, you know, our standard practice for confining space entry requires at least two to three people. um one person going down in the hole, one person manage keeping an eye on the person that's in the hole and then the third is usually the you know we have like a a ticket system that they fill out um and kind of the manager that's kind of keeping an eye on everything else. So with the idea of utilizing a drone, you could cut your people, you know, you could do this with a single person or even two. to potentially reducing your personnel um by a third to a half uh for confined space entry. **[15:50]** >> All right. And I think one of the common challenges I think we we've spoken about you know how it was done uh manually before uh how humans used to do it and now how drones are using it. Uh but what about other challenges such as you know people who are building their drone program you've taken about 2 years to build it from the ground up yourself. uh what were the other challenges like when it comes to regulations? I'm sure that's that's something that the audience would uh really benefit from as well because they're a common pain point, right? So, how did you navigate? >> Yeah. >> Yeah. The reg the regulations, you know, in in each uh you know, the EU has theirs. Um the FAA is is United States, they have their own regulations around it. um in each airspace, each country has its kind of uh guiding regulations and it's always it's a constantly evolving topic. So staying on top um of those rules as they're evolving now with uh beyond visual line of sight **[16:54]** drones. There's a you know new proposed rule. There's a part 108 in the US that's going to be coming out in the next year and it it is going to completely change how u beyond visual line of sight drones are operated and regulated. So there's kind of two pieces. You got to keep you got to keep tabs on the um regulatory piece and then you also have to keep your tabs on your technology piece cuz the technology is also evolving just as fast. Um sometimes it's actually evolving faster than the regulatory um side of it. But it's kind of an everchanging field. So when you're going kind of into this program, I would definitely identify um stakeholders and then try to I would say go and elicit some um volunteer support. Find out who's interested in the program and say, "Hey, you know, I see you've got a really good background in electrical systems. Um we you you're going to need an electrician to install these systems." um or you're you have a network **[17:56]** specialist that's been really into drones or they have you know they kind of on the side got their license then you know I would approach those individuals. Um the legal side of things too to kind of find there's I think in every profession there's probably somebody who has some interest in drones or has heard about drones. So, I would identify those people and kind of bring them into your program because you're going to need you'll need support um from all kind of facets of the organization. Um it's a lot to stay on top of uh for just one individual. So, um a lot of the initiative starting these initiatives, you don't really have a personnel to start with. You kind of kind of have to seek them out in the organization to see who's interested. [snorts] Right. So that's that's very interesting. You know, rallying up all the other internal teams to be a part of the broader program. So they they can contribute, you know, to the program success. That's that's really important. Uh you know, again, **[19:00]** it's contingent to the program success as well. Uh, I know that I've been like asking a couple of back and forth questions, but if there's anything specific that you'd like to share, feel free to do that and drive it as well. >> Okay. Yeah. Let's see. The I think the big one that kind of just touch base on is I think a big one for the audience is, you know, somebody people just starting in the drill program. I know a lot of that's probably where a lot of people are at right now. There's I think there's very there are more organizations that are growing and scaling but I feel the majority of the audience probably today is probably like hey I'm just I'm trying to start this program where do I start from? Um so that tip that I I shared earlier about um kind of really identifying stakeholders and interested parties um and make a list of those for staying on top of regulatory uh items and and technology pieces. Um [snorts] **[20:01]** something uh I'm trying to think some other tidbits. There was the you know return on investment stuff too. Um that I've I've kind of built a internal business case because I'm the surveyor and we're also trying to have our own fleet of drones internally. Um so with surveying uh some interesting stuff that I found was uh for a 25 acre site um that will take a drone roughly a day uh a day to map and process um that will take a a regular traditional survey crew that can be anywhere from three to six five days to map that site. So a return on investment can be a one to three or a one to five ratio. Um and then on addition, you know, while you're doing these mapping projects, you might catch a security breach or some other options. So there's other value streams kind of plugging into that at the same time as well. >> Um so kind of looking at all the **[21:03]** different value streams that these systems can provide. I think that was a good pitch for us uh when we went to the Chiefs again was that multi-m missission um stacking you know showing that there's not just these systems don't just do one thing it's they can do many things at the same time too um with you know have AI enabled uh video monitoring in the background too so you can actually have second eyes watching the video feed um so from a security standpoint that's a that's a big win Right. I think very interesting take you know like again figuring out the initial use cases figuring out the value that they provide and the ROI. So you know if something was being done in 6 days and now it just takes one day to get that done. Again that's that's a clear uh ROI plus obviously it's critical infrastructure u you know safety is involved as well. So all of those things just sort of add keep **[22:04]** adding on to that entire value and ROI uh narrative. >> Yeah. And it's most of even with just one value stream I was finding that most of the ROIs were within the year uh fiscal year. So you were able you know it was pretty easy. You're saying hey if we can provide value for one of these streams most of those were at least on the survey side of things that was all calculating to be less than a year. um which is not you know when you're budgeting that's a really I think that's a big win for finance to say hey we can we're this isn't you're not I'm not telling you a fivey year sixy year return it's like within the year we can get a return on the on the investment so and then you start piling on those other um value streams and all of a sudden that that [clears throat] ROI starts ticking down really fast as times as timeline so >> right and I think one thing that a lot of people would be interested in um learning from you is let's say you gave an example right how do you rally up other **[23:05]** teams sort of like source talent to get the job done uh for the drone program >> how did you manage the change you know people are not really open to let's say adopting new technology there is always some kind of friction >> so uh did you track any success metrics for the drone program just to show everybody how it's working >> yeah yeah yeah That's uh that's a good Yes, I did. So, for the pilot program, we ended up we had about six pilots that we identified that got licensed and went through our our program to become a remote pilot in command. Um and then I created a form to kind of track their user experience with the drone. Um, I felt that was important because, >> you know, like any new tool, um, if it doesn't work or doesn't provide the utility that the operator has perceived that it should do, then they're not going to use it. So, we really didn't want to get this invested in this system and then all of a sudden **[24:08]** everybody's like, "This system doesn't work. I'm not using it." And, um, while Ian and I, who was my my partner in kind of getting this program going, we were very excited about it. We realized that uh not every operator is going to share the same interests as us and obviously we're we're a little biased um on our reports. So I created this form, a user experience form. Um so that might be a good useful tool if you're starting a pilot program um to kind of track track the utility of the of the technology. >> Wow. Perfect. So u let's also talk about let's say okay you've built the program you can see the success internally initial use cases >> how have you been thinking about let's say scaling it to you know multiple sites again you opened the right uh you know slide >> yeah so the to scale um so the the next stage for that was to you know update Our **[25:14]** so our original waiver from the FA was only for a single site. Um we just received our nationwide RA waiver back in July and part of that component was creating a site selection matrix. Um and in that site selection matrix it's really similar. You know you're doing a assessing risk um and FA they call ALR acceptable level of risk. So you're identifying airspace, ground stuff, communications. These are kind of if you can see this table, those are all kind of the major categories. And each under each category there's um you know five to 10 specifications and we kind of make a score of that. Um and then we score each site and I have it reviewed by my team to see if they agree with me. Um and then we set a threshold. So then now we identify we have properties that we want to install these systems at. Um so essentially we repeat this process to identify new sites as that **[26:17]** we want to install these systems that meet the our our standards and regulatory compliances. And then there's also a testing component to that. So um this is this image over here on the right is one from FlytBase. They've got a a heat map that can monitor and manage your RF link. Um, and this has been a really useful tool for us to test the connectivity around a site. Um, so we'll just fly missions around the entire area. Um, and then you can get back uh the connectivity strength in a heat map. Um, so we kind of built this into our uh concept of operations document. Um so that's it kind of this is kind of how the reg from the regulatory side how we scale and then um from the actual physical side uh is more budgeting and and getting money getting more money for all these systems. Um so we're we're pretty close and we're ready to to kind of scale this up. So we're right now we're in the process of identifying sites. We also have uh let's see if I have a **[27:24]** image. You can kind of see here we as we vehicle mounted the dock 3 um for testing. So we can kind of go to each site >> um and actually do tests at different locations. Um that was kind of a nice thing, a bonus of the doc 3 is it does have the ability to vehicle mount. Um, and so we kind of go to different sites and and launch um a launch from these different locations to test if they're optimal >> um from the kind of like the physical side of it. >> Wow. I never thought that you know figuring out if you want to scale to a new site obviously you know budget value ROI that narrative is there but like your approach is very nuanced right the the scoring for each site based on multiple uh factors to figure out you know is it even worth scaling to that site from a regulatory standpoint or you know other environmental network conditions as well that's that's amazing **[28:26]** >> right So um I think we're almost just 1 minute uh on time. Um we mostly covered everything. Um I'll do a quick recap. Uh uh but what we can also do is we can take up a couple. So uh let me just check in the chat section. So okay, one question is what's one mistake or challenge that you think another utility could avoid because you've been through it? Um I think engaging uh just again we made the mist we didn't engage I think enough stakeholders early enough um because now that this program is going I'm getting I'm getting a lot of I'm just getting I'm daily getting peppered with emails asking uh about you know what can we do with this drone or um and now everybody's interested and we're we're kind of hitting the limitations of our certain specific **[29:27]** scale. It's a good good problem to have, but if we had maybe engaged more stakeholders earlier, then we could have gotten a little more budget um for that program, >> right? And like you said, it's a good program to have and you know, hopefully u you're able to figure out how to cater to these different use cases. Uh okay, one final question and then we can do a quick quick recap and get your thoughts. So we're all talking about physical AI, right? In the future, where do you see a utility like Denver Water uh leveraging like uh let's say AI and how do you imagine AI taking not taking over but basically augmenting what you already do, >> right? Yeah. And um so the big part I think is on the inspection side of things. Um as you can see up here on the screen, this is just one uh quarter mile section of a long concrete canal that spans about 15 miles. Um **[30:31]** and part of that is we have we have to monitor and keep an eye on some of the um the cracking and and some of the spalling. And um so we over here have been kind of playing with an AI tool to kind of identify um cracks and spalding and and defects in the concrete. [snorts] Um and it's by no means a perfect you know it gets us we say 70 to 80% of the way but it really zeros in on areas that have uh need more attention from the engineer. So instead of >> you know it doesn't replace the engineer or the inspector but what it does is it actually focuses the attention on areas uh that are probably need more attention instead of looking at this giant you know concrete structure and saying oh being feeling overwhelmed. The AI can kind of assist us in inspections to say hey you should probably look over here. Um, so it kind of I would say uh what do you say narrows narrows the scope for the uh **[31:32]** professional um uh engineer and and inspector so they can have be more efficient with their time um in inspecting these structures. I think that's a very interesting take because that's what we seeing in the industry as well like the drones autonomous drones and docs they help you collect a lot of data and then AI obviously is able to process all of it and like flag the key areas that you actually need to go out and inspect rather than you know telling you to inspect all of the flagging a lot of things uh that you should not be putting your attention to for example. >> Yep. Exactly. [snorts] Yeah. These the scale of a lot of these structures is just a little overwhelming for you know even a team of two or three people to consider inspecting >> right >> y >> right so I think uh Brian we've do you have any final thoughts I'll just summarize the session in let's say 30 seconds like you know we started with the why on why even Denmark waters thought **[32:35]** about creating a drone program how you and your team figured out the early use cases, how did you rally up the different stakeholders and teams to get their buy in uh figured out you know the operational change build the program showed the value uh the method for scaling you know I'm sure you have your lessons and have figured out okay if I want to replicate this through 10 or 20 of my sites this is the playbook that I need to follow and we've also you know uh done sort of a deep dive into what the future would look like uh you know in the age of AI and autonomous operations. So really insightful you know discussion I got to learn a lot. I'm sure the audience did too. Uh any final thoughts from your end uh before we close the session. Um yeah, I think that you know at Denver Water, our our drone program is really about extending the reach of our workforce. As infrastructure systems become more complex **[33:37]** and workforces become more limited, uh regulatoratory expectations continue to grow. Um these tools are are an excellent tools to manage our critical infrastructure. Um, and really kind of excited to see where this where these programs go and all the utility that, you know, hopefully others uh can learn from this presentation or from this Q&A and um look forward to sharing more lessons learned in the future. >> Right. All right. Thank you so much, Brian. Um, you know, for coming on stage and sharing your story. Like I said, very insightful. Lot of great learnings here. And thank you to the audience as well for joining in and you know uh listening to all the insights that Brian had to share. Uh well this was it for this session. We have a couple of more uh lined up as well. Uh so yeah feel free to go ahead and join the other sessions as well and thanks once again Brian for sharing your story >> know **[34:39]** what the future looks like for your operations. >> Awesome. Thanks Neil. Thanks everyone. Bye-bye.