# Turning Fleet Maintenance and Compliance into One AI-Native, Predictive System - Event: NestGen ’26 - Published: October 2, 2026 - Length: 25:28 - Watch: https://nestgen.org/library/watch/R0z4wL7qwsY/ > 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 As drone fleets grow, maintenance logs, battery health, flight records, and regulatory paperwork tend to spread across spreadsheets and separate tools, and problems surface only after they cause downtime or an audit finding. In this NestGen session, FlytBase will share how Prediq brings fleet maintenance and compliance together in one AI-native, predictive system, and how teams can move from reacting to failures and paperwork to anticipating them. ## Transcript **[0:00]** certified. Are we following all the regulations? Most of you in this room have already figured out the flying part. If not most all of you, right? And you've chosen your device, like your hardware. You've picked out your software, amazing. And you've got the integration working, or at least you're on your way to getting it working. And that's the hard part, and it's done. That's great. This session is about the part that comes next, keeping all of it compliant, maintained, certified once it scales. Because that's where we see programs slow down, and it's rarely for the reasons people expect. Right? So, it's fine. It's not the real agenda. Just quickly get started. So, let me start with a pattern we see across customers. Scaling usually breaks in one of the two situations, right? Situation one, you have built the operations, hardware is in place, the software is integrated, workflows are running, your team is flying every day, everything works. That's great. And then someone asks, "What happens when **[1:03]** we are at 50 sites? Who's keeping track of licenses across all of them? When was the When was the 12th dock last serviced?" And you realize that operations scaled, but the compliance behind it really didn't, right? Or you have situation number two. You're about to scale. The POC's done. It went well, and the expansion plan is signed off, right? Or maybe you're making the plan. And the plan covers the dock, the drones, connectivity, again, software, all of the things I mentioned before. But what it usually doesn't cover is a tool for compliance and hardware maintenance. It's not that anyone forgot. It's that a POC at scale, you at a POC scale, you didn't really need one, but now you actually do. Right? In both cases, compliance turns out to be the biggest road block, and it's not one thing. It depends on your pilots as well. Are they certified? Are they current? It depends on your hardware. Is it maintained? Is it trustworthy? And it depends on it on your software. Can it prove all that when someone asks? **[2:05]** If any of these three is under invested, the program stalls. And I And I'd add one more thing. Two years ago, none of this was a big deal, right? You had a few drones, you flew it by hand, you had a generic logbook somewhere, right? Uh or a spreadsheet was maybe enough for some people. The The reason it's it's a big deal now is that operations have changed shape. Docks fly on their own at night uh at sites nobody visits for weeks. Uh the paperwork used to grow with the number of pilots. Now it grows with the number of docks and there's um no one on site to actually do all of this, right? So, scale brings new questions. And I want to be specific about them because compliance it it it's a big word, right? That that I think hides a lot of meanings. Let's say at 50 sites, how do you make sure everything stays compliant? Licenses, waivers, flight rules. And not at one site, at every site. And all the time. How do you make sure every piece of hardware is maintained? I don't just mean the drone, I mean the **[3:07]** dock, the batteries, maybe the RTK relay attached to it, maybe all the other payloads that you have on top of the drone. So, or maybe the propellers of the drone, everything around them, right? And how do you make sure every pilot is always certified, the trainings are up to date, right? And the people the right people are cleared to fly on the right sites. Now, tools for this exist today. I want to say that clearly. There are good fleet management and compliance products on the markets and many of you use them. We are even, you know, integrated with them. But here's the question I actually want to ask. Are they built for dock drones? Not really. Most of them were designed for a different kind of operation. One where there is a pilot in the field with a controller and the tool records of the pilot being There's a That's a reasonable design. It just isn't the operation you are running anymore. Think about what dock operations actually means. There's no pilot on site to sync a log. The dock itself is a piece of hardware with its own **[4:07]** maintenance schedule. And in most tools, it doesn't even exist as an asset, right? There's a base station, a stand, a payload, a rotation of batteries, and the whole thing runs around the clock. A tool built around a person and a controller can record a dock flight. It can't understand a dock operation. And that's what we're here to see. Let's make this even more concrete, right? With three types of operators, three types of personas. Uh these are completely hypothetical. I mean, drawn from inspiration with some of the deployments I've done. But, if you're running docks today, docks today, I suspect you're one of these three, and one of these things probably if you relate with, if not all of them. Right? First, the program lead. You're rolling out docks across 50 sites, and some of them are uh you know, sensitive sites, uh maybe even like offshore platforms, right? Tank farms, sites near controlled airspace. Your job is to make this a program, not a set of, you know, POCs. **[5:12]** And the question you keep coming back to is, "How do I stay compliant across 50 sites? How many pilots do I need? Who's certified for which site? Who's due for, you know, retraining?" And then you think about this as a multinational conglomerate that's spread across the globe. These sites are spread across the globe, across countries with their own jurisdictions, with their own regulations, and that's Even I'm tired just thinking about it, right? Second, the service provider. You won a contract a run of 100 docks across solar farms, six in the morning to be running from 6:00 in the morning to 6:00 in the evening and you know, of course, there are SLAs, right? And then the question you ask is, how do I prove we actually flew safely? Do the other reports generated fast enough if something goes wrong? Is everything actually logged or do I again have to make a manual reports every single time about this? You know? And the third operator **[6:14]** is, you know, flying 24/7. You're maybe in a remote operating center, around the clock, fully remote, nobody on site and your question is a different one. How do I prove we actually, you know, how do I keep every dog in battery working? Uh I mean, if you have, let's say, another 30 dogs or 40 dogs, how do I make sure every device, everything is maintained? Now, of course, these these are questions that is applicable for everyone. I'm just trying to paint a picture over here for here over here. Notice what's at stake in each case. For the program lead, one compliance gap at one site can put the whole program under scrutiny. For the service provider, an incident without a clean record is a regulatory problem, not just an operational one, right? And for the 24/7 operator, a dog that's down for a week is a contract conversation. So, three different operations, three different questions. Hold on to them because the second half of this talk is going to answer each one of **[7:15]** them. Okay? So, how are these actually teams actually handling this today? Uh Honestly, there's a separate team for every part of the operation. You have one for the weather, uh for maintenance, you maybe have spreadsheets or a shared drive, uh you have one for compliance, a folder of licenses and waivers, one for pilots and one for the reports you send to the regulators and this lots of screens. Right? And we see this in real operation centers. Someone's sitting in front of five or six windows and a chat thread on the side asking when was Doc 12 last maintained. For each of these tools and for each of these tools it Well, it's that they are fine on their own. The problem is that what they have in common. They were built for manual flying. They assume a person is on site flying and updating the records afterwards. With Doc, that person doesn't actually exist, not on site at least, right? With five Docs, this actually works. Someone can hold **[8:16]** it together. With 100 Docs, it falls apart. Not because anyone is careless, but because every one of those screens depends on a human keeping it current. And there are now hundreds of Docs and no humans on site. And here's what that looks like on a Monday morning. The ops manager opens the weather tool to see if it's safe to fly. Opens the maintenance sheet to check nothing's overdue. Opens the compliance folder to make sure no waiver expired over the weekend. Opens the pilot roster to see who's on shift and whether their certification is current. And then if anything went wrong, opens the reporting template and starts reconstructing what happened from four different sources. And all of this for your first mission of the week. And this is even before a single mission has actually been flown, right? And that's the gap. Not a missing feature. It's a it's a missing category. Right? And that is what we built. Predict. Predict is **[9:18]** uh a predictive fleet maintenance and compliance tool built specifically for Doc operations. I want want to underline something specifically. This isn't a general fleet tool with a Doc feature added on. It was designed from the start around how a Doc ecosystem works. The Doc, the drone, the battery, the payloads, the base station, the stand it sits on, any of the additional hardware or the kits that are part of it, right? And the pilot that's on site behind each one of them. And it sits inside FlytBase. That matters and I'll I'll show you why in in just a second. >> Most UAS teams track flights in spreadsheets, maintenance in shared drives, and compliance in folders no one can find during an audit. As fleets scale, these manual gaps become the single biggest risk to the operation. Introducing Predik, AI-native predictive maintenance tool for smart and efficient fleet maintenance and compliance. Predik replaces manual workflows with a connected live hardware ecosystem. At the **[10:22]** moment a drone lands, critical operational data like flight duration, mission logs, battery cycle health, and dock status are captured automatically in real time. Nothing is left for the pilot to fill in. Readiness is tracked in real time. Monitor pilot cumulative hours and certification renewals automatically. Use the schedule module to plan recurring inspections and coordinate missions based on actual asset availability. Maintenance is no longer a guessing game. Predik comes pre-configured with manufacturer-recommended service intervals, making maintenance tracking far more reliable. It continuously monitors per-drone flight hours and automatically triggers [music] alerts as aircraft approach critical service thresholds, helping teams prevent missed maintenance and unexpected downtime. Centralize your licenses, waivers, and SOPs in the compliance center with automated expiration alerts. Log and track safety incidents to maintain an immutable timestamped audit trail for every asset in your fleet. For audit-ready records, context matters. **[11:23]** Predik logs robust weather conditions, including wind speed and temperature at flight altitude, ensuring your records reflect the actual environment of the mission. Finally, transform this data into intelligence. Generate compliance-ready PDFs or structured CSVs for regulators and insurers in seconds. What used to take days of reconstruction now takes a single click. >> And this is the core of it, right? Over here. First thing, automatic logs. As soon as a drone lands, its log is pulled from FlytBase and linked into the dock, the drone, the battery, and the pilot. All of it is mapped. You don't have to go out of your way to register each of these things and manually input all of the details. No, it's there in the log. It's there over here, right? And all of this will happen within seconds of the drone landing. The second, no manual mapping, as I said. This is the part that changes everything **[12:23]** downstream. In most tools, someone has to tag flights, assign them to pilots, link batteries to aircraft, and fix whatever didn't sync. And that's the work that quietly grows faster than the fleet. In Predict, it happens on its own. Because Predict already knows what FlytBase knows, right? And the third one, analytics on top. Maintenance, compliance, and reporting all run on that data. So, they're accurate because the data underneath them is accurate. And it's accurate because nobody had to enter it. The only things a person still enters are the things only a person can. For example, your license expiry or certificate or waivers. Everything else comes from the core of flow operations. And I want to be clear about why built into FlytBase matters. Because it's easy to hear it as a sales line. It's It's not about convenience. It's about where the truth lives. FlytBase already knows which dock is launched, which drone flew, which battery was in it, what was on shift, what weather What was the weather conditions **[13:25]** like? Any tools sitting outside that has to be told. Skydio does not have to be told, right? And that's the whole difference. And it's a difference between a record you maintain and a and a record that actually maintains itself. And that has value to it. So, what actually happens when a drone lands, right? So, let me walk you through it step by step. Very briefly, it just as an overview, right? One, the drone lands, the mission ends at the dock. Two, the log is pulled. It comes straight from FlytBase. There is no export, no upload, no sync button. Three, this is the one that matters. It's auto-linked dock, drone, battery, pilot. Everything is part of your log. So, every flight is attached to every asset it touched. And that's pulled as well. And the fourth, analytics update, maintenance intervals move forward, battery cycles tick up, compliance records refresh, reports are always current, right? No setup, no tagging, no cleanup. **[14:31]** If you ever spent a Friday afternoon reconciling flight logs, that specific line is for you. Because I know people who who, you know, spend a week who fly for a week, maybe 2 weeks, and then over the weekend hundreds of flights are then, you know, synced up. A lot of work, right? So, three main questions that I think, you know, are the core of everything. One, how do I keep every dock and battery working? Now, let's go back to our three operators. Question one from the service providers. The second one of them, part one, is hardware maintenance, right? And this is all of this is set up in minutes. Because DJI's maintenance manuals are actually built into Skydio. When you add a dock or a drone, its maintenance schedule already there, recommended intervals, checklists, replacements. You don't actually have to build it from scratch because now there is a software that can build it for you or have it already built for you, right? Next, one single **[15:34]** calendar. This one calendar is it has data from all of your, you know, scheduled missions. Yeah, so all of your devices operating hours, it has all the upcoming maintenance, it has the pilot schedule on it. So you don't actually have to put in like three different, you know, spreadsheets and check out three different timelines and then try to overlap them. No, everything is just like in one place and the view updates itself because it is driven by flights. Uh technicians are also logged. Jobs get assigned to technicians and their work is logged against the asset. So the next time someone asks, "When was the Doc 12 last serviced and who did it?" The answer is right there. And this is not just about the drones. Uh it's about payloads, RTK base stations, dock stands, kits, everything on site that has maintenance requirement, it is tracked alongside that because I know that some of us have found ants in our docks, right? And this can help with that. So let me put that back in the service provider's day. **[16:38]** A dock on the solar farm 2 hours away processes its service interval, predicts, raises the task before it's overdue, overdue, assigns it to the technician covering the region, and gives them the checklist from the manual. The technician does the work, logs it, and the dock's history is updated. Nobody had to remember, nobody had to check a spreadsheet, and for the service provider, that means fewer docks are down and SLAs are always met. Battery analytics, right? Part two of the same question is batteries and I want to spend a moment here because the number of uh numbers surprise people, right? Take Take that 100-dock service provider. Around 10 flights a day per dock, a battery is good for a fleet 400 cycles. Do the math. That's around 900 batteries a year. More if you're in extreme heat conditions. Even more with like other bad weather. So, that's not a maintenance detail. That's a budget line. That's an SNA **[17:41]** risk. And most teams find about find out about it when one failed or like find out about it one failed mission at a time when things are already up and going and then you're like, "Oh, this happened." So, PredictFlight every battery individually, its health, its cycle count, and which dock and drone uh it is in. And even if it is switched between docks and drones. So, the historical logging of it is also there. It flags weak or aging batteries early before they fail on missions and it tells you when replacements are due. So, you can plan them instead of, you know, discovering them later on. Now, question two. From the operator is flying 24/7, how do I smoothly and report a fast if something's actually wrong? First, law compliance check. Every flight is checked against your operation operating limit. Did it stay under the maximum altitude? Did it stay inside the approved area? That's recorded with flights automatically, not constructed later on. And let's say something does break **[18:42]** compliance, a flag is raised, right? And you see it with your reports. Weather at flight altitude. Uh this one is specific to docks, right? Because uh ground-level weather does not actually tell you what's happening at, let's say, 120 m above ground level. PredictFlight checks conditions at that altitude and the drone actually flies and keeps that with the flight records. You get, you know, fly with caution or no fly or good to fly. You get these kind of warnings way ahead in time. And So, when someone asked whether it was safe to fly at 3:00 a.m. in the morning, you can actually show them what that was like at different altitudes. Shift and incident reporting. Shifts and incidents are logged against the flight, the drone, and the dock, and the pilot, everything an investig- investigator would ask for is already connected. So, again, let's paint a picture. Let's see why this actually matters. When something goes wrong in a manual operation, there's a pilot who can tell you what happened. In a dock operation at 3:00 in the morning, there isn't. The record is the **[19:44]** only witness, and if the record is scattered across five tools, an incident report takes days and still has gaps. If a record has built automatically at the moment of the flight, it takes minutes and it's complete. For 24/7 operator, that means evidence ready in minutes and not really in days. Question three, right? Uh how do I actually stay compliant across 50 sites? And this is more about, you know, your licenses, your certificates, your waivers, uh all your pilot records, all of your reporting, they are all housed in one place. But, the interesting part about is on this part, because I know there are tools out there that solve it. It's that you see all these expiries, you see all of that information on that one single calendar view, right? So, it's audit ready at any time, even if it's you're talking about five sites, but actually talking about 50 sites over here. You just have to open that one calendar view. You see your operations, you see your maintenance, you see your compliance uh dates, everything. **[20:48]** And I think that provides a lot of value, right? So, we have different operators, we all have the same problem. As you scale up operations, you need tools that are actually built for docks, right? And that's what we are, you know, working towards. So, three very different operators, a program made across 50 sites, a service provider with 100 docks and SLAs, an operator flying around the clock, different scale, different industries, different questions. But, underneath it all, it's the same problem. And if there's one thing I'd like you to take away, it's this. When you plan a dock program, you plan the hardware, the software, and the connectivity. Plan the compliance and maintenance layer at the same time with the same importance. And plan it for docks and not for a pilot with a controller. Because the operator will scale. The question is whether the paperwork scales with it or against it. And that's the thought I actually want to leave you with. Thank you very much. Uh I hope **[21:51]** this was informative. Uh I would love to see I I see there are some questions over here. So, uh first question, we already use Airdata, why actually switch? So, uh well, we are integrated with Airdata as well. We don't have to switch. It all comes down to what kind of an operation you want to render uh run into and, right? Uh for dock operations, as I've just said for the past 20 minutes, there's a whole bunch of things, a whole bunch of reasons. This this one single pane of glass is why I would recommend the switch. But, the decision yours is yours, the operations are yours, and at the end of the day you can decide. Uh next. How does automatic mapping How does mapping happen automatically? Okay, I'm I'm I'm assuming we're talking about, you know, the pilot docks drone mapping them across uh the software. Uh so, when you you must know that every time you fly uh any device and concept plans, a flight log is created. And that flight **[22:56]** log has all kinds of information. It's It's rich with information. Not only does it have all of your flight details, it has hardware details, it has, you know, uh pilot details, it has moment-to-moment details about what happened when. So, we take those details, we take those logs, and then we actually just map them out, and then the software maps that maps it out automatically. I I hope that answers the question. >> Uh perfect. And let's see this one. I have another one. Does it work with non-DJI hardware? >> Yeah, so as you know, we are a hardware agnostic software. We are actually, you know, integrated with other hardwares like Pixhawk and Starting, right? So, of course, it's 100% possible to integrate, let's say, maintenance manuals from these other hardware providers and get them in here as well. So, you don't actually have to worry about it being DJI specific. Because we **[23:58]** want to support all drone operations, and that's what Predict is actually built for. >> Uh we have Jeffrey is asking, "Where does the data get stored? Is it all cloud, or is there a local hosting option?" >> So, it depends on your deployment. If you want uh if you have a cloud deployment, then yes, it does get stored in uh cloud, right? And whoop, sorry. Yeah, it does get stored in cloud. Uh this can be, you know, FlytBase on its own cloud servers, but if you have your own private cloud, we can do it in there as well. But let's say you have an on-premise deployment, maybe it's on like an edge machine, or it's on your own VM. Uh we can have uh this uh living within your ecosystem as well. So, completely flexible. I hope that answers your question, Jeffrey. Uh I don't think we have any more questions. It's been a pleasure being with you over here, and I hope you guys learned a thing or two and I look forward to connecting **[25:00]** with you as well and seeing where this takes us in the future. If you have any more questions about this, of course, you can reach out to FlytBase and or me directly. Thank you very much. Have a good evening or day and I'll see you at the next session. Bye-bye.