# Panel Discussion: Blueprint for Running an Enterprise Drone Program - Event: NestGen ’26 - Published: October 2, 2026 - Length: 49:04 - Watch: https://nestgen.org/library/watch/3NmpSQCwTwQ/ > 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 In this NestGen '26 panel, Jean from Skeye, Julie from ABOT, and Achal from FlytBase discuss what it actually takes to build and scale an enterprise drone program, from the first dock to multi-country operations. The conversation covers the full lifecycle. Jean shares how Skeye went from two flights a week to fifty a day at a refinery in the port of Rotterdam, starting with tank roof inspections for safety and expanding into perimeter checks, piping integrity, and gauge reading. Julie brings the perspective of deploying across multiple regulatory environments, from France to Belgium to West Africa, where the rules change at every border even within the EU. The discussion gets into why some programs scale while others stall after a successful pilot, the role of trust and how long it takes to build with a customer who has zero tolerance for incidents, working inside ATEX zones, cross-border regulatory fragmentation, and why the technology can be standardized far more easily than the operational and compliance framework around it. ## Transcript **[0:00]** joining in. Uh it's time for the first panel discussion at Destin. We had amazing session so far. We still have half of the event left. You'll have the keynote soon and there'll be more sessions. We'll be starting the oil and gas track uh in dedication track that is coming up along with containing security, physical AI showcase and the other vertical that we are going to cover. But this is a great starter session because this would uh give us that preface to a lot of things which are covering and plus it will help us go deeper into the enterprise drone program. So let's uh get started uh with the session. So of course the uh blueprint the session topic is blueprint for uh running an enterprise drone program and we'll go deeper uh in with our panel experts. Uh we have today John and Julie. So first let me quickly introduce myself and I'll hand it over for them to introduce. So I'm of **[1:03]** course uh co-founder and director of business development at FlytBase. Been in FlytBase more than 10 years uh being drone doing drones all my life and I am responsible for working with all lovely partners customers all across the globe and we have amazing panel uh on our board today. Okay. So we have Jean joining us uh from Netherlands and we have Julie's joining from France. Uh I'll hand over mic to them to have a quick introduction. So Jean why don't you go first and quickly introduce yourself and Sky. >> Well thanks for having me Ael and nice to be here. Uh my name is uh Sean Kilus, director Leos at Sky in Rotterdam near in Hanord actually near Roterdam and uh quite recently joined the company as per the first of January having a great time and doing wonderful things. More uh more on that later I I expect in this panel discussion. **[2:04]** >> Awesome. Julio to you. Hi everyone, I'm Julie Gil uh from Habot Company. Uh thank you Asha Flybe for having inviting me. Uh we are about on the French market uh more than 10 years now to provide technology solution in drone and robotics to a whole B2B um activities in France. >> Awesome. Thank you so much both of you for joining today. uh let's close the presentation and let's move to talking because this is all about panel and discussing about various topics. So uh let's uh get rolling. So the first story or the part that I want to cover is there are of course everybody's deploying drone docks. Uh things are getting started but the companies who are adopting this the programs stall. I mean it we have seen companies probably buying autonomous **[3:05]** drone dogs one year ago and they still haven't scaled. They are is still seeing that problem. They're not seeing the next things how to go ahead. Maybe it's ROI. Maybe it's something else. So let's open up that question to the panel and both the experts. So Julie why why don't we go first to you. I think the customer first comes to you always if once they hear about drone docs right if they are in say France they would come to you and say hey I'm looking for this technology they would come and see the demo what have you seen in last 12 months like if they have bought one >> what has stopped them to scale what has stalled the program >> right and is it just France or are you seeing even more countries uh uh in Europe or in even other French speaking countries which you serve so would have to understand your perspective. >> Yes, for sure. Uh a lot of more and more companies are interesting to adopt drone autonomous but there is a single gap uh between to think about it and to adopt it. **[4:09]** So the biggest difference is that the customer who scale don't start with a drone they start with the operation. In fact, this is the way of mine with changing the project that stole um a refer technology driving. They bought a dog but and they say okay let's see what we can do with it but finally the project they scale have already identified the operational problem and the person who holds who it internally the workflow that will change and what success look like. I I would say that there is a very early warning sign saying that if there is three month after adoption, nobody has really taken ownership of the program internally and and already it's it's a risk. A successful program needs an operational owner, not just a drone enthusiast. They have to understand this is the world solution, not only um funny drone tools that this is the way that we can move forward on this, >> right? **[5:11]** So what I'm hearing is maybe previously a drone pilot could just purchase a manual drone and they are just flying it. But when they are looking at drone dogs, it needs to be a structured program approach. There needs to be a program manager, right? Who who needs to take because it's not just about flying a drone. There are so many other components and we spoke about that in the keynote earlier morning and we'll we'll have the replay again all the different layers. But that brings me to John because John, you have already deployed these things with oil and gas company majors like Shell, right? >> What justified the program and they have kind of scaled already, right? They've been using it and like what were the use cases that unlocked it? There would be lot of things that had to overcome. So help us a little bit in that journey and what what what was your learning? What can audience learn from this? Yeah, what what we've seen is that uh uh um before I joined Sky obviously that there was a a need to inspect uh the tank roofs of the tank farms in Rotterdam **[6:16]** in the port of Rotterdam. And basically what what was required was uh an an automated inspection of tank roofs whether there was water on it and whether they were uh operating properly. And uh the base case in in in our scenario was uh we want to have people out of hazardous uh uh positions and out of climbing icy stairs to uh check on on on the roofs in in winter time. And that that's how we started off with only two flights per week. Uh now up to 50 flights per day. Uh so that that's that's how it scaled. What happened is that the initial flights were at a at a pretty high altitude, far away from structures, far away from the tanks, far away from from piping, etc. And um uh in in in due course uh the customer realized that there's more to gain from having the drones in the air uh and to not only **[7:19]** inspect uh for instance the the tank roofs but also to check the perimeter also to check whether people are were intruding uh the the area but also are my is my piping still intact. uh can we finally can we uh actually capture the the gauges and the valves and can we do something with that information and that's basically how basically how it literally took off from just inspecting roofs to doing operators but but from the the basic assumption that u we need to have we do not want people uh to be in hazardous environ ments that was the principle in the in in in our case in in the port of them. >> Got it. So what I'm hearing is generally it's of course about cost safety exposure uptime. What you're seeing is safety exposure dominated more right. Is that the correct understanding? >> Yeah, that that was that was the the the the entry **[8:23]** level assumption. And I think that if we now look at the amount of data that we uh that we capture and that we deliver that's being processed then uh it it has in addition also caused an uptime of the of the factory of the refinery ultimately uh which I think is where not the the human factor is the big game but the uh the the financial factor is the big game. >> Got it. So cost is also playing it. U digging down here like why did this program worked like what could have gone wrong if and it would have stalled but it hasn't stalled and it hasly deployed has been deployed like if you look at look connect the dot backwards what made it work were that these are the only things that that you said or there were any particular thing that you think because of this they actually decided to scale and now you have like from two flights to 50 flights a day. Yeah, I I think I what was paramount in in in our example is uh that you start small and **[9:28]** first gain trust because everybody is is anxious about hey is this safe uh will the drones uh cause incidents or not and luckily we can say that in all the flights that we've done which are are over 10,000 only in in in beef loss capacity in the port of Rotterdam we did not have one uh one actual incident. Um, so we we started really small. We uh uh helped people understand what we were doing. We were on site during the first flights. We are always on site during new flights or new mission creations just to explain to people what are we doing, why are we doing what precautions have we taken uh to prevent anything from going wrong. uh and what are the actual advantages of having such a versatile system as a drone in a dock deployed in in in an industrial area. >> Got it. Connecting this again back Julie to your part where you said you need a **[10:29]** program manager but what is expected of program manager? Is it like the right hardware specs? They need to define the use case who owns it internally. Uh how how does it matter? Have you seen any kind of examples where uh like in Fr France or any other countries that you are deploying uh what what exactly is something that has went right or wrong for them to get started with this program? >> Um regarding regulation you mean? Yeah, regulations it could be again safety uh cost the pointers but even if somebody has a program manager like one is of course we agreed we said you need a program manager once you have a program manager like what are the things uh that person he or she need to kind of take care of is it the use case that needs to be picked up like John said it's about it was about safety exposure first right where I kind of everything yeah **[11:31]** >> for Yes, the the regulation in is the main topic even um even though there is the European framework now which has sure harmonized everything. Um finally the adoption locally is not uh enough of um is not obvious. So um yes it depends uh of the national authority type of operation sites uh and everything has to be taken into consideration uh to get the adoption uh also and um this is the global solution which needs to be compliant to also the environment of the local company [snorts] >> right >> so yes this is everything >> yeah yeah I think that's a bridge to the the next topic that we had about compliance and John I'll bring it to you because uh this was something which was also solved by sky right and that kind of enabled in a way and I'm talking about the **[12:32]** first is the be loss part right >> without which like you could have a P with a line of sight operator but this technology can't even take off no pun intended there right you cannot have drones uh there and so and specifically for your use case it was a refined binary. So there's one upnot of safety and what you need to do. It was not a simple security missions. We had lot of security sessions today uh where you are doing parameter search. I I also spoke about how to build security program but it comes to this complex use cases and compliance. Let's just first talk about loss and then we can touch upon things related to uh other parts. So would love to understand what went through uh you can of course give us more insight about your luke your luc as well and help us understand how was that journey and how is that important uh for to scale >> what what we've seen in the in the port of Rotterdam case is that um the LUC we obtained for for BV loss uh was **[13:35]** a great help in establishing clarity with local local authorities um to uh gain access to the the actually to the skies uh of of Rotterdam uh also operating in the CTR of Rotterdam airport which was an additional uh factor to to be taken into account. um that I think uh took us three or four months only to uh to obtain a regulatory approval with the proper authorities and explaining to them what we were uh doing there, what we intended to do and how we would do it as safe as possible. Uh and then obviously the customer also has uh high standards as as as regards uh uh safety, flight safety, safety of the site, uh instructions how to operate on the site, permits to be obtained before uh before flying and I think that took almost the same amount of time in only in preparation and only after that **[14:41]** preparatory work uh we could start uh slowly in uh uh entering uh uh the facility in uh explaining to the people in the facility what we came what we came to do and uh how we would do do it and how we would do it as safely as we do it up till up till up till now basically and uh you can imagine that on a on a on a facility like a refinery uh you have uh explosive areas uh a tech zones uh that you need to take into account and that you also have to make sure that the drone will never pass uh through or across an an Apex area also not when a return to home is uh is initiated by the system or by circumstances. Uh and that's um you need to find out you need to explain you need to obviously program your drone in a in a proper manner and also instruct your own personnel but **[15:42]** also the staff of the of the customer what happens in case of an of an emergency and luckily as as I indicated we never had any incident so far but that is because of all the precautionary measures we take from a technical perspective, from a procedural perspective, but also from a communication perspective with uh with with the customer. Uh lately, we've we've adjusted our own procedures uh to be able to uh fly closer to objects and lower to uh to the above ground level. Um uh we've added in our procedure that we were all always to use a lighter on top of our M4 drones. uh um if we were to fly close to uh to to objects and that's obviously technology advancing. Uh we make we make use of it but it's all also always checking um the uh the optimization of of of of risk because we have a very low risk **[16:49]** appetite and I think our customers customer even has a lower risk appetite. So we can't have anything go wrong. Um uh and we do we do yeah we go the extra mile to prevent anything from from happening. >> So far so good. Yeah. >> So so you touched of course the B law you also touched about the compliance of HSC and what changes you made. Was there anything that surprised you during that process? uh anything that that you never thought of that hey this something needs to be taken care of >> uh in in in search of uh of in in terms of compliance. >> Yeah, in terms of compliance. Yeah. You were getting into this. >> Well, the the the the the fun thing is that we can from a from a flight safety perspective, we can do more than the customer would would want us to do. So uh we have more confidence in uh the capabilities of the technology right now than is acceptable uh and it's **[17:49]** also and that's a matter of trust. Trust you have to win and it takes a very long time to win it. Uh and if you lose it, it's gone in an instant. Uh so um yes, we recently had internal discussions. We can do uh missions closer to the objects maybe even inside factories u but uh that's uh that that's a bridge too far literally. Yeah. >> Got it. Got it. And did the expertise that you had with manual drones because Sky has been in the business for a long and you have flown drones manually. Did that certainly help getting the LU or it helps getting into that process or it's a very different world altogether? >> No, no, no. It it it definitely helps. Uh our company was established in 2012 and our founder is still with us and he's sort of a celebrity in the Dutch drone world. Uh and that I think that helps a I'm sure it helps. Yeah. >> Got it. Got it. Right. **[18:52]** Moving this to Julie you you actually shared about this that EU harmonized story right that hey things are harmonized there's one be loss uh John is talking about how they are doing their operations we talk from the other side right let's talk about France or some other French speaking countries nearby right where how how is it kind of working out and probably you might also have something in Africa right where uh lot of stuff we have deployed and seen that how is it different what's going on with regulations there and what is your observation? um it's um yes as I said previously it's not harmonized yet even there is a European ESF framework now so for example in France the basic regul principle um France Belgium are the same but the compet the the the national authorities could in gave their own interpretation then rules **[19:54]** um national rules indeed more than ESF framework. So for example in France we are dealing with the DGAC so um like national authorities and so the processes as Sora for example is um used regularly um but it's taking time to get SAR for example in France it's taking 6 months you see and John just talked about LU LU is not approved in France yes we cannot use LU you see um at that time but uh I hope it will change even we are in Europe both of us so there is still some um different stage uh according the country um we are working on that to uh with local authorities to to harmonize everything for example Switzerland which is uh EU member states there are the own aviation **[20:55]** authorities also and the horn regularly framework. So it's also um different type of authorization that in France uh we are working also in Africa, western Africa mainly uh which is quite also very different. There is no sor authorization at that time. uh it's easier for sure but the security um uh the security framework is important too to take into uh consideration there is not only uh one um uh one authorities for many countries. So each countries has its own local authorities to convince and to work with to get the authorization. uh you may have to deal with the local operator requirements, import and custom procedures, local pilot qualification which is different with specific flight um authorization. So for an enterprise operating across several country, it's not **[22:01]** easier because yes um you have to um pay attention to very local authorities even if there is a national European for example your friend point. So this is the the key point the technology can be standardized much more easily than the regulatory and operation deployment. So when customer tell us um we already have an operation approved uh we want to do the same in Belgium with Zander Africa we have to say okay wait we just need to study the local country to see if we can how we can um display deploy the dog there because it won't be the same rules as we did in France necessary and we see where Jean is using sometimes you see in in rotam so it's not very far from France, but we cannot use exactly the same uh the same way of working. **[23:01]** >> I I fully recognize what you're saying, Julie. It's a it's being interpreted differently in in almost every individual country. Recently, we did a demo in in Norway. Uh we send a cross border notification under the LUC and within two days we got an answer of the Norwegian authorities. Uh you're welcome and good luck. So [laughter] that that's what happens as well. But it it's not uh always the same reaction. Yeah. >> Right. >> Yes. >> It's science of the growing industry. Right. So this this wasn't even the case before. There's discussions like I remember earlier generally people would say hey how would I deploy this? This is just a toy. I can't even use it. But at least now there are signs that we we see that kind of things happening. And uh yeah, I'm sure it will come for like how Julie is saying that it should come for France as well and some more countries. U I would like to switch gears uh to more one of the risk that **[24:02]** also comes into building this program is the supply chain availability. So Julie you are an expert about manages lot of products and solutions right? U I of course know from uh American friends where they face a lot of challenges. If somebody's in Europe today, right, and they're looking to get hundreds of these, has there been effect in supply chain uh with all the geopolitical issues or things going on like how how do they protect if say they're going with DJI doc or DJI doc 3, right? And they want to scale and they're looking for 50s or hundreds and everything is going good, they they want to do that like are there any challenges there? Are there any lead times uh things which which you face? >> Yes. Um actually one of the most underestimated aspect of a drone program this is the yes the trip and haveability. In fact when a customer buy a dog they just don't buy a piece of hardware and and that's it. They are buying availability a world **[25:07]** solutions need working 24 hours. So availability depends uh on much more than the manufacturer repairs time also. But we assume uh the the the repair times in the after sales and um they they need to um my recommendation for for for the customer is to um uh define the service level around the dog the need because okay we provide um the solution the J and then okay fine but then we it's needed to be there is some redundancy see necessarily some repair parts availability and so on replacement units so on so they have to think about the whole solution not think about okay it's a drone I just need a drone and that's it now it they are thinking a whole solution and um this is the way we are there to provide them the the level of the services **[26:11]** they need yes so they don't buy just a hardware but a wall availability Got it. And in terms of Yeah, sorry. You're on. >> Yeah. No, I I I recognize what Julia is saying, but I al also have a a funny anecdote because uh um recently I got a request for drone in the box systems for a battery factory uh for for a car manufacturer and they were actually they they've uh um approached the the the request from the IT perspective fully. So they thought they'd buy a bit of hardware that we would train a couple of pilots and that we would arrange an LU for the company and then that we would fly out again and leave them as as is. And uh obviously you can try and do that but that's not a sustainable system from maintenance perspective from training perspective from keeping up with the regulations perspective and it was uh it was remarkable and I I I wonder **[27:18]** what happened in the end there. That's uh because it is what Julie says. It's a it's a system that you uh want to acquire or purchase uh with a a minimal uh amount of availability uh every day uh the whole year through because it is an important part of your operations. And that's uh yeah that's that's funny that you >> yeah you you shouldn't approach it as a as a piece of hardware that the IT department can buy and then use. Yeah, >> makes sense. And Julie on the parallel track when it comes to the geopolitical exposure, right? So one is of course getting the hardware and have it in stock and you are able to repair it. We have seen lot of requests for sovereign software right specifically when we are dealing with lot of French customers or similar countries the requests have been always sovereign on prem. So what's the story? What's what's going on? And is that is that the legit requirement everybody has as they're **[28:21]** deploying this hardware? >> Yes. What we see in in the French enterprise market uh and also public administration that is not simply a stop or go situation. Customers are becoming much more conscious of data cyber security supply chain resilience geopolitical exposure and so on. But in many case they are not saying we don't want the GI. Okay. They are asking a different question, different way. They they they ask for how do we can we reduce our dependency and control the layers that matters to us. Um that can mean sering that hosting software layers integration with existing system alternative sensor local capabilities or simply having a clear content plan. So um I think the market is moving from pure hardware decision towards an ecosystem and risk management decision. They need to be uh reassure **[29:22]** um to to to be confident about the solution. They are not um saying okay no no DJI they just want to have some solution effective that make them secure. >> Got it. So that adds one layer of more complexity right because one is we're talking about hardware itself is complex don't look at that just hardware but now you're getting into onremise and servers and installations and so on right so then it becomes a quite uh complex or what we call enterprise uh kind of deployment that comes into the picture >> right so >> John on the same lines if if you were looking at somebody like shell like how do you manage availability how much spare capacity do you have to keep uh for say a single dog fails uh right or how to maintain that it doesn't break or you need to replace together are there any anecdotes or any um ideas that you have seen being implemented there? >> Yeah. Yeah. What what we um **[30:24]** obviously availability is a is is a big thing um uh which uh leads us to always have uh uh u spare uh docks, complete docks, spare drones, obviously uh spare engines, spare uh props. But um most importantly uh we put our dogs on on trailers with a with a router on it and with uh uh spare capacity in communications and uh in electricity supply. Um and we always have uh one or more trailers available. Uh but we also have one or more uh docks in even in the in the carton boxes still available to uh to replace them quickly. Uh and u yes availab availability is is is a big thing. We uh and obviously most likely Julie will do the same. We uh supply all the uh the dock systems, all the routers with with double SIM **[31:26]** cards with Starlink if required based on on the on the location where the trailer is. Uh the dock then has a or sorry the drone then has a an extra SIM card inside based on two different uh network providers in the Netherlands. And then obviously back in the office we have uh our fixed line. We have our uh 4G 5G line and the Starlink and we have uh extra electricity supply. Lately we we bought an old-fashioned diesel generator to be able to continue works even if if power fails uh uh from from from the grid in the Netherlands. So it's it's it's uh precautionary measure after precautionary measure after precautionary measure to be able to to maintain service well until the bitter end. Yeah. Yeah. >> Got it. Got it. >> Great. So that brings us to the last topic before we get into **[32:29]** a crossfire and a lightning round. So the last bit is of course scaling that drone program right from getting one to one dog to creating that program that we talking about uh Julie started the concept talking about program manager. So Jean maybe continuing with you what changes operationally between one dock and 10 plus doc staffing ratios per pilot shift models what do you centralize what are the learnings there? Yeah, that's it's it's a it's a good and fair question, especially uh uh regarding the amount of security and surveillance solutions we've seen come by uh today at at Nesten. Um the we are allowed one pilot is allowed to fly two dogs at the same time uh within our regulations within our beefy laws regulations and within our LUC. Um and the funny thing is we don't do it a lot because uh within our own procedures **[33:29]** we've established guidelines that only simple missions and missions that are flown more than 100 times by a pilot uh can be performed together. And we don't have that many simple missions because you can imagine that uh a mission flying high over tank roofs is an easy mission. you stay away a fair fair distance from the object. But if we fly to obtain data from a refinery, we fly low and we fly close to the uh to the objects. And basically those flights do not permit for scalability in the sense that one pilot can fly multiple missions at the same time. Uh because under Dutch regulations are under Dutch interpretation of YASA regulations. Uh the pilot should follow visually uh the BV loss operating operation that is that is continuing. So uh we can do two missions, one pilot **[34:36]** at best currently. >> Okay, >> got it. And one is of course pilots and drones. One another thing that scales is data. Uh right. So the kind of data you're collecting. So when it comes to sky and shell's example, where does the data reside, who consumes it, which workflow did it change, uh and who acts on the output? >> Yeah. Well, we we are the collectors of data are basically our our drones are they are the the the taxi show for driving around and writing down basically writing down the numbers by taking pictures. Uh all those pictures are assembled and are transferred to a partner of of our customer that applies machine vision uh to it and makes an interpretation and provides digital data back to our customer that then incorporates it into into their operating systems. uh which uh which means that uh if we don't deliver we have people on uh on **[35:38]** on the telephone immediately uh because it's a vital part of of the information infrastructure of the customer. Uh so it's uh it's crucial information uh assembled and and basically brought together by by drones and then an IT intermediate. Yeah. >> Got it. And Jimmy over to you. Now the customer has done the P they're using this looking to scale. How who do they call? How do they work? Who as a program manager if somebody's listening here what are the step they should take and uh how do they plan to scale from here? >> Yeah. Um they shouldn't have to choose between contacting distributor integrator software company for one side you can manage those relationships. um and separately um the the inter interface uh is uh is could be the problem. So the responsible **[36:39]** um when adopt uh DJI do works um he is as to have only one contact uh to manage everything and uh to coordinate um the supplier of of different um thing and the team their teams. Um so um the the the ideally yes the that's why we provide uh we provide the whole world services hardware software integration training and so on. So in this way it's easier for them to scale. >> Got it. So a one-stop shop is is actually they should look forward to because then it's anyway a very complex system and then uh you as a partner are already closely working with uh the as integration part as a software part and then could closely work with all the other partners and uh they don't have to actually pick up **[37:43]** three different call phones to fix things right they can just call you and then you take care of it. Awesome. So let's uh start the crossfire and the lightning round. So this would be questions open to both of you, right? So say if if somebody in France, Netherlands wants to start with five sites tomorrow, right? Uh how would you propose? What do you recommend to them? Uh right. Would would you pro propose them to buy hardware software? Would you propose them to outsources? Maybe this is the debate we can have between the two and see what how what would your recommendation would be. >> Yeah. Yeah. Well, I I would first want to understand what the purpose is of of calling me, calling us uh and and check what what what the main goal is. Is it taking people out of hazardous positions? Is it only focused on on efficiency on cost levels? Um but mostly if you uh uh are considering drone in the box **[38:45]** systems you need to have uh in my opinion you need to have repetitive missions. If you don't have repetition then you should either go for a vlo solution or consider other measures to be to be taken. Um that that that would be my first my first take on on on that on that request. Got it. Jules, what what would you recommend? >> Yes, just as um yes, the same. Of course at the beginning we need to understand well um the purpose before everything else and then we will bring to the customer all the um all the tools that they need what kind of hardware software and also networking and and and the network and um what kind of system um they need to have to be compliant to their own system and to and to their requests. **[39:49]** Um, yes, this is the way. >> Got it. And follow up or flipping it to the question maybe John to you that what's the most expensive mistake, right? You have watched customer make maybe they didn't come to you or or they even were there but they did not agree to that part and and they you they actually made that expensive. Has been there a case where you have seen that happening? >> Well, I I only work for for for nine months now for for Sky and during my time there were no mistakes. So that's that's that's no no just kidding. Just kidding. No, but um what I think that is the most costly mistake that we could uh make with a customer is uh to push too hard too fast. So, uh, what what I said before, uh, you need to, uh, explain it not only to the the the people that are signing the contract, but also to the people on on the premises, the staff that are walk walking **[40:50]** around that are actually seeing the the drones fly. They need need to understand why they fly, what the what the aim is of the of the drone operation, and why it is conducted in a safe way. And if you push too fast, if you want to go too hard, too fast, then uh you will it'll it'll bite you. You will you will be haunted by the consequences. And that is people that don't want you to deploy drones or that say it's it's it's more dangerous than uh putting putting staff in in hazardous areas. And then then you're losing. I think you lose you lose trust. You lose faith in the in the in the product and the proposition. And uh ultimately you will not be able to uh give the customer the benefits that are attached to deploying deops. Yeah. >> Got it. And to you it could be both side either be we can make a mistake to the customer or customer is making a mistake. Have you seen any kind of examples or any **[41:51]** anecdotes that has happened? >> Uh yes it's happened. Uh obviously because it's only the beginning for only 18 24 months the kids. So I think one of the most expensive mistake is underestimating everything that comes after the purchase order. Customer often have a very detailed business case for the drone itself but not for training, maintenance, spare parts, regulatory work, data integration, internal change management and so on. So those cost don't necessarily appear on day one um because we can this is the first time so there is some new cost appears during the process. So the real cost of a drone program is the cost of operating um over months even years. So we have to take the whole chain into uh consideration and that's why uh for both parts we underestimating some some cost. **[42:53]** >> Got it. Great. Thank you. Uh we would be running out of time just couple of minutes left. So few more questions and we can take one or two from the audience. So one is like this everybody's of course calling this drone program but it is more than the drone program. So what's the right frame? uh maybe one of you can take that is it and does it live inside uh security team asset integrity team HSSE who owns the budget like what's the right way to approach and imagine end customer is listening to us right and end users how do they think about it >> yeah it's it's an interesting question uh and I think it's it's it's very good to realize who is benefiting from the the drone program I think that in in case of operator rounds like like like we perform in the port throttam um it is operations that should benefit uh it's operations that should also have it budgeted and the funny thing is that operation **[43:56]** drones are versatile so you can use them for uh ERP missions in case of emergencies rescue people um uh but uh if uh ERP is fully depending on the budget of operations uh they can get in trouble. Uh if oper if operations for instance isn't interested in thermal cameras uh whereas ERP requires them to uh find people that fell in into the water or something then uh then you have then you have a problem. So I think it's it's it's important to distinguish between the purpose, the benefit and the the the the people that included in the budget. It should be as as clear as possible. >> Got it. Got it. Awesome. That brings us to the fortune telling or prediction couple of question. Maybe Julie I'll give you a sentence you can fill in the blanks. **[44:56]** So the the the sentence is say in next 24 months what would enterprise drone programs would be able to do routinely? What do you proact predict uh in next 24 months and what are things you still think won't be able to reliably do. So that eventually would answer what everybody needs to focus on for next 24 months. In the next 24 months, enterprise drone programs will routinely perform autonomous repetitive inspection across multiple site where the data automatically entering enterprise workflow. Um but um they still want replace human judgment in a complex inspection and decision making. But the technology will automate more of the collection. the value will increasely come uh from what humans do with the information. So um yes I think for the future it's a wonderful tools will help everyone. **[45:58]** >> Got it. Thank you. Great. Uh yeah correctly said. And last one I would uh to Jean like one thing every enterprise should do in first 90 days if they're looking to get started now. I I think they should uh um from from scratch measure measure measure and measure and then uh find out how much can be gained by deploying drone in a box systems. Yeah, >> got it. Great. Thank you so much, John. Julie, that was a great conversation. The time just flew by, right? We're already overrunning the time but there was great conversation on various topics. We covered uh the entire building scaling the program from the compliance part to supply chain issues to scaling the programs at multiple fronts. Uh I would just take maybe probably one question if we have any Shi if you have any questions if you can pop that up. Oh, we have few of them. Maybe **[47:00]** let's take uh one from Ricardo. Uh where Ricardo asks that data capture. Okay. Yeah. The most enterprise program gets stuck in pilot program because they optimize for right frequency rather than bottom line process replacement. At what specific unit economic threshold or metric did your program transition from an R&D innovation budget into an operational line item funded directly by BEu? Maybe John this is for you. You already said last question, last thing was ROI. So >> anything share with you? Well, the funny thing is obviously that the customer doesn't share all his metrics uh with with us, but uh I expect uh that uh the quality of the measurements combined with the reduction in uh in in costs uh for for gaining the data was a clear uh clear reason for customers to uh to expand the program and to incorporate it in their in their in their **[48:03]** enterprise operations. systems. So, uh yeah, I I can't say it was 100,000, a million or 100 million because I honestly don't know. >> Got it. Sounds good. Uh we have few more questions but sorry I would not be able to take it. We need to be respectful of the time. Uh and if John you're available, we would have networking sessions and audience can join there as well and drop them the question uh right there. I again thank you both uh for taking out time um busy schedule to talk about uh very important aspect of uh building enterprise programs and uh contributing to the community. Any last thoughts before we drop off? >> Well, thanks thanks for having us. I suppose it was great to be here. >> Yes, thank you. Thank you so much. Thank you everyone. >> Okay, thank you. Excellent. Byebye. Bye. >> Bye. Bye. Bye.