Otovo has found ways to really reduce their dependencies and usage of SaaS software. As a reminder, this is pretty close into Q2 earnings season. They're in a quiet period, technically. We'll not talk about their financials. They announced on their Q1 call that they are exploring a potential U.S. listing. We can't talk through that. Let's just focus the discussion on how they are changing things with their deep usage of AI. With that, note that this webinar is live and being recorded. Here are some of my disclosures. Rohit has some as well. I'm going to pause and let him show his screen, and then we'll jump over to John and Jack with their deck.
Cool. Thank you, Phil. While I pull up my PowerPoint, welcome, Otovo team. Looking forward to seeing what you have done with all the AI tokens that you have spent over the last six to nine months, and how that's changing the business that you do. Again, would be a proof point of what, from the tech investor standpoint, what we see as extremely powerful tailwinds and headwinds affecting more than 200 companies in my comp sheet. This webinar is live and being recorded. These are the mandatory disclosures. One Mississippi, two Mississippi. Thank you. With that, thank you again, Phil, for doing this. Thank you, John and Jack. I can't make it who's the son, who's the father from the pictures here, but feel free to jump in. Again, looking forward to the discussion.
Thank you, Rohit. All right. Let's get started. We've got a slide deck. Then we'll go into Endurance here at the right time here. We have our own disclaimers. I think everything's been disclaimed at this point in time. We can go to the first slide here. First of all, name's John Berger. I'm CEO of Otovo. Jack is Head of Software and AI for us at Otovo, has done really amazing things. What I want to get to and thought that this would be something very interesting to our shareholders, both the U.S., Europe, global. It's really exciting. It's regardless of industry, no matter what it is. We'll get into and talk a little bit about Otovo before I turn it over to Jack. I hear constantly about artificial intelligence. Is it real? How do we do it?
Is it really just using a ChatGPT, or is it something more than that? Of course, we've all heard about SaaS apocalypse, what's happening with the software, what is real, what is not. We've seen the software companies, including this week, defend themselves and say no. We also saw IBM just last week have its worst trading day in its entire history because of concerns around AI and what we're about to show you. Exciting. I think there's a lot of implications, Rohit said, in the world. This is regardless of solar, generators, power, whatever it is. I think you're going to get a real treat here today. First of all, on the slide, let's go through and Endurance, if you look at what Endurance is and what it means to us, just the highlights here.
We have one operating system. We have one confirmed margin engine. It's repeatable, allowing repeatable M&A. That's our highlights. This is what we're going to tell you. These are the summations about why Endurance is important to Otovo. What is Otovo? Let's take a look at that. Otovo is essentially an AI-enabled service company roll-up. It started with a platform company of Otovo that was in the solar space. It was doing origination. We shut that down. Now we're moving and have been buying companies at a breakneck pace. Eight acquisitions. We just closed SST. In fact, Jack's in that SunSystem Technology. Jack's in the Northeast U.S. doing the integration. It is within three weeks as we close that, I think July 1st or 2nd, Jack.
We are closing on, as we announced previously, we're working on closing on Green Panel acquisition. Really bolster our European operations tremendously. We're looking forward to that. As we move towards being profitable, real exciting in terms of our plans and what we're doing. We are really, really executing. Very happy and looking forward to the earnings call. I've got to tell you that. I'll leave it there. When you look at our company, we have over 1.8 million. This does not count, by the way, SST and Green Panel. We have 30,000 customers. These are Q1 numbers. These are all sanitized, so you can imagine this number is larger as we report. We see just $5 million in cost savings, which is tremendous. That's not including SST, it's not including Green Panel. It's not including on a forward basis.
This is an annualized number. It would be repeatable every single year. It's growing materially. We're going to talk about that as we get into what the benefits of Endurance are. We did this basically with a very small amount of people. In fact, majority of these adds, if not almost all of them, next to Jack, was really since May. We've done this with very small numbers. You've heard a lot about that AI. We can tell you that that's real. Again, what we're going to show you is we're going to show you reality. We're going to show you instead of talking about AI and talking about buzzwords, we're going to show you what we've done, and we'll show you the numbers.
We've talked about this on the Q1 call. We're going to be able to show even more real cold, hard cash and how much has been saved. When you look at in terms of what Otovo is in the next slide here, let's start first, how do we make money? We make money through service. We are a service company. What does that mean? We're rolling trucks and fixing and techs, and we're fixing generators, solar, batteries, EV chargers, and load managers. We're doing that in Europe, U.S. and soon to be in the Middle East as well. We are the largest behind the meter, as soon as we close Green Panel, we anticipate that, behind the meter power company or power service company in the world. We service both residential and commercial.
We make money through memberships, signing up for service, and then paying us on a per month basis. Repairs, fixing solar systems, batteries, generators, make money on that. Upgrade sales, so think batteries and so forth. Retail power and VPP, so that's aggregating power and selling it back in the wholesale system where appropriate, which is all of Europe. Right now, we're doing that in Texas. That's how we make money. What is Otovo? It's a crossroads here. It's three things, power, and when I look at power, you look at the space, and for two different reasons, both Europe and the United States, we all know this, has gone through a tremendous downturn in solar. There's a lot of good service companies available out there. They're subscale, but they're good at very good valuations. We've been taking advantage of that.
It's gone much better than we thought. We've been very acquisitive, as I said earlier. Post even the merger of Otovo, we've done eight acquisitions, and we're on a tear. We have a big backlog. These are profitable businesses that we're coming in and further cutting, and Endurance in AI is a big part of those cost cuts and pushing that operating cost, that overhead cost lower. The second thing is service. Service has been terrible in this industry. There's reasons why a lot of people assume that an installer is the same as a service company. I built scale businesses in both areas, and I will tell you, installers are good, and it's a project construction business measured in weeks and months, and service is measured in hours and days, and that's a logistics business. It's two very different businesses.
Also, there's a lot of bankruptcies that have been happening out there, particularly in the contractor space. That's left a lot of manufacturers of inverters, batteries, generators, stranded with customers that are looking for service and looking to get their power back on. That has been exacerbated by the huge build-out that we've seen over the last 10 plus years. It's now the chickens, so to speak, are coming home to roost. Somebody's got to go out there and service these customers, both residential and commercial, and that's Otovo. We're seeing a lot of demand from a lot of B2B customers to come in for the service. Last, that's the purpose of our call today, AI. Got to say, be honest as always, I fell into this. It's just the right timing.
When you look at a tremendous amount of our costs as an industry, it's actually SaaS software. It's coming in and having to spend the money. Then when you look at the back office, we're talking dispatchers, schedulers, marketing, finance, procurement. All of this is a lot of money. A lot of money goes against affordability in solar and behind the meter generations, particularly distributed that doesn't have the scale as front of the meter does. This technology, AI, and what we're doing with it, so exciting. We're attacking that cost and we're doing it. We're not going to talk about doing it. We've done it, and we're going to show you that here in just a minute. As I'll turn the call over to Jack. Jack, I'll say yes, there is a relationship here as everybody I think understands and sees.
Very proud what he's been able to do. I'm too old for the technology really understand. I can get by. You really need to be, I think, in your 20s and maybe even younger to really be a savant at this, and Jack certainly is. Done a tremendous and very exciting. Haven't seen anything like this out there, period, in the economy. I think this is, like I said, it's a real treat. I'll turn the call over to Jack and walk you through Endurance and what it means for Otovo. Jack?
Hello, everyone. Thank you for coming today. Here what we've built out in Otovo is the Endurance operating system. What does that mean on a bunch of different cost bases here? We have marketing, sales, supply chain operations all brought into one central operating system for Otovo. When we started out making Endurance, the idea and the impetus behind it was not necessarily to make it for AI and just shove AI into everything. In fact, as a company, obviously we do a lot of AI now, but we wanted to make the best CRM as a basis. Of course, with AI and using a lot less developers than before but not using AI properly in the operating system. I believe that we've done that, and over the past few weeks, we've now injected AI deeply into each aspect of these four pillars as well.
What you'll end up seeing with the Endurance platform is you're going to have lower acquisition costs, much higher throughput of work orders and appointments that we can send our technicians out on, much faster M&A integration. Our average integration time is three weeks. Massive revenue growth that is independent of our overhead growth. As we grow our revenue, usually with service businesses, you have a significant problem with the back office increasingly exponentially growing as you do that. That's not so for Endurance. Today when we do our live demo, we're going to show off the marketing system. We have a full marketing email, SMS, CSAT, CMS system, everything like that you would normally pay for or break off into a separate system is all into Endurance.
Same thing with sales, commission tracking pulling up AI sentiment, site reports, everything like that, supply chain, the inventory management, taking requests from partners all the way through the process and automatically ordering these parts all the way through. Of course, operations on automatic scheduling and load balancing our technicians in order to increase our margins on those.
Hey, Jack, before you go to the output, can you say as of 1st of December 2025, if you had to run your business, just walk through like people who don't understand how your business works, maybe just okay, you get an order, you do this, you do that, you do that. What was the time? I just want to get a good quantification of how much time are you saving and how much money are you saving before and after, if you can just pinpoint that.
Yeah. I'll go over this more when I go through the proper platform itself. The idea here is let's just take SST, as I'm here in New Jersey right now, integrating them into the system. What do they do when they integrate?
Just to be clear, SST is not solid state transformer.
Oh, yeah. It's SunSystem Technology, just for everybody else on the call. That's one of our recent acquisitions. What they were doing on a significant amount of overhead basis was even if they're handed the lead, no marketing, no nothing, just B2B, they're taking that lead, they're getting it in their inbox, they email it. They go manually put it into their CRM system. Okay, well, now they got to get somebody to go call them. They call them on a personal cell phone. They're like, "Okay, well, let me just put a increment each time I call them." The same person has to call them three times because they require three contact attempts. You don't know what was said in the previous call, so you just log that you called them. It's not in a system.
Of course, you can get a SaaS to get it all in your system, then your costs increase once again. Then everything is a two-monitor system. You got Google Maps in one, and you got your open jobs in the other, just praying and hoping you set up the correct route for your 10 technicians in the area. It's very, very difficult. It's a very, very difficult math problem in general, which is what we've optimized entirely into Endurance. There's a lot more to it, and I'll show you as we get in here.
That problem, that huge back-office problem, following up the customer, sending them the site report, saying, "Hey, this is what we did on site," making sure that they've collected payment, all of that is what the SaaS, in order to even complete those tasks, is already going to be about a 10% even margin. On top of the labor savings, we will do much more than 20%-30% margin impact. We're bringing that more in as we're bringing more AI into the system. What we've seen is about $2.3 million in SaaS and at least $1.7 million in SaaS support that we've cut on an annualized basis over the past year. That SaaS support is probably underdone, but the old Otovo Europe had already undergone massive cuts before we took them on.
In reality, the engineering team that built Otovo Cloud, which is their old legacy system, was nearing 100 people at the peak in order to actually develop the system. They were down to just less than 10 once we acquired them due to cuts, but that was not a build new features team, that was a maintenance team. An additional million so far in staffing. That's cutting out dispatchers, cutting out office space for those dispatchers, cutting out a variety of different back-office people that are having to do accounting work. We're doing a lot of integration with our NetSuite and Ramp providers that is done much, much quicker than we would've been able to do before, and much cheaper.
Okay, fantastic. Also maybe a later question, maybe John. Things that you are doing better, faster, cheaper, much better. That's one bucket, and I would love to, towards the end of this call, get your feeling of things that you couldn't dream of doing and now you're able to do. For example, you couldn't sit on your phone and get a car in two minutes because now you have Uber there. 10 years ago, you couldn't dream of that. You now have 5 million songs in Spotify, for example. Things of that nature that you couldn't dream of a year back, John, like, "Hey, by the way, now we have this technology and now I can do it." Maybe I can acquire 30 companies in a day or whatever that thing could be.
I would love to think of like you have 30 minutes to think of, be as creative as possible as a public company CEO. I know as a sell-side analyst asking that is not a good question, but think of that. I want you to stretch your imagination and think now that you know what this technology is capable of doing. Yeah, if I can do that'd be nice and cool. Anyway, go ahead with your demo. Just 30 minutes ahead of you.
Yeah. I've got an example demo for y'all set up on exactly that. Here we have the SaaS graveyard. We've cut a total of about 56 to 100-
Hey, Jack. Sorry to interrupt. Can you go back to the prior slide? Sorry.
Just to put a finer point, on the right side here, guys, the margins are going from 15%-25%, that's a 10% 1,000 basis point improvement, right? That's kind of what this $5 million of cost savings translates to. There's potentially more beyond that as well. Just want to make sure that's very well understood. Sorry, Jack.
Thanks.
Yeah.
No, that's good. That's a good point, Phil. This is what we've realized. We do see some other firms out there, General Catalyst, Thrive, Founders, doing a roll-up strategy in their own service verticals, and they're saying they're seeing at least 10% on just the SaaS alone, and then the additional 20%-30%. We can verify that, at least for what we've seen in these companies. We're stripping out at least 10% of the cost, and that goes directly to the bottom line.
Yep. An Endurance license is $0 for us. Here's our SaaS graveyard. Got some fun Otovo colors in here. This is just a subset that we've put on here. But you can see some heavy hitters here, including AWS, which is kind of interesting. Usually, even people that build their own systems still rely on an AWS to host the underlying operating system itself. But here at Otovo, we've actually built our own server architecture by hand. We actually have our own servers, we own our infrastructure, we have multi-region failover, we have oversight and monitoring over the whole system, and we have an always-on team because we're a global company, but we also have an always-on team in the form of our AI engineering that we've also built out behind the scenes. We have an always-on AI engineer as well.
Yeah.
Okay. I'm going to switch over to Endurance proper now, I believe.
This is the real system here, is what you built. What I would also say is just from my perspective as CEO and goes to a little bit of what Rohit asked me to think about, this is unfathomable to me, and I think many other CEOs that have seen this about how much has been done. This would take, first of all, we would again procure to the tune of tens of millions of dollars a year, these software pieces. You saw the graveyard. To build that software that they then sold to us for tens of millions of dollars a year, you're spending more CapEx, like couple of hundred million dollars building out what you want for your business. This is astounding. The speed at which you can change is just mind-blowing.
It's because you own the entire stack and you're pulling it in as you keep going because you can. The technology is so powerful and you own your entire stack. There's a lot you can do with it. Wait, we'll get into here later about the agents, the agentic AI creating the agents. This is their own playground. You don't have to go in and have a lot of blockers because you have all this SaaS stack of 50 to 100 SaaS companies. This is not just a cost play, but this is also enablement of becoming a truly agentic AI company in whatever your business is. Jack
Yep
Do you want to walk through?
This is Endurance. If you're wondering what the guy in the boat is, Otto is our company mascot in all of our marketing material and things like that. Endurance was named after Shackleton's boat, which if you're not familiar, Shackleton was an adventurer and one of his very, very famous boats was the Endurance. The idea was to kind of endure any challenge, both in the service industry and through the changes with AI. Here I'm going to show you probably, honestly kind of a random one that I ended up pulling up for this system. I wanted to show the power of Endurance. We're a global company. We have a lot of different languages; we have a lot of different resources across the globe that it's very hard to manage.
You'll see pretty clearly here where Endurance becomes really powerful in order to be able to manage all of that. First of all, everything in here is translatable. I can work on this despite not knowing the language. There's a lot of languages in Europe, I don't know most of them. This is immediately extremely helpful here. The reason I picked this case for this customer was this is a customer that emailed us saying that they had a battery issue from one of our installs. I wanted to pull up here. Iris is our AI assistant that we've just deployed recently in the past few weeks. Iris found an email chain from FoxESS, which is a battery and inverter manufacturer, that said, this client said, "Hey, I'm having a problem. My battery is not working.
It's saying it's not charging correctly. Anything like that. Iris is handling this case right here. It's handling it in lockstep with our human assistants. Iris can handle it and then hand it off and then take it back over very seamlessly all the way through the system. In this case, I particularly like this because FoxESS emailed us saying, "Hey, by the way, this person, their battery's approved." They didn't really tell us exactly who it was.
They did tell us the serial number. They did say, "Hey, this is the serial number that's approved for a free replacement." Iris went into the system, was trying to find out who this applied to, found the install pictures from this person's install on the project that we installed the battery for them and found the serial numbers off the install pictures, linked it back over to this case and said, "Oh, hey, Mr. Customer, FoxESS has approved this for you." You can see the translation right here. The manufacturer's decided to replace the main unit. The address provided will be used to ship the replacement module. Once it's been delivered, it will be possible to determine the next service steps. That is a perfect example of Iris and our AI inside the system doing a great job in handling the customer's query.
This all happens within five minutes, which it's extremely difficult to do this. This is a huge problem in the service industry, being able to respond on time and promptly to these customers, especially as we get in new information like this that is quite pressing to the customer. The customer has a real issue here. I wanted to highlight that at first, also show you all the different systems that we have in Endurance that compound upon each other. As you saw before, I'm actually going to switch from out of our EU system into our U.S. system here just quickly. You can see all of our monitoring alerts coming into the dashboard, quite a few of these systems are in active failure. Not all of these are our customers.
We have purchase lists and different things like that that allow us to see a lot of different, from old bankrupt companies and acquisitions and things like that allow us to see quite a lot of monitoring alerts and act on them for these customers. We brought that all into our system here. That automatically brings that into our lead intake for the system. As we're going through here, we have all the calls, texting, emails. We have the dialer system that in order for you to dial out, we have our quote system being able to quote automatically through the system, along with a bunch of different items going through the system as we're intaking these. Something I want to show off here as well.
You'll see we are intaking these work orders from our B2B partners and other things like that and actually quoting everything with AI. This is not a simple process by any means. Quoting in general in the service industry is extremely difficult, in large part because it is effectively a complete system design. You need to get the conduit. You need to get the exact length. You need to get the exact type of conduit. It needs to be electrical resistant, whatever else it is. It's extremely difficult. And what we've found is that we have, after only a two-week trial period on this system, we've had about a 95% exact match rate with our best estimator at our team.
Now we are going to start taking these work orders all the way through the process and immediately scheduling them for the customer because we know how much they're going to cost for the customer, for the B2B customer or the B2C customer immediately. You'll see here quote revisions and anything that AI built on here. It also brings in the parts required, and then we'll actually bring the parts all the way through. As you can see here, the AI has reserved these parts required for this system. Here's our OT numbers, which is just our Otovo inventory numbers, and that is going to bring this copper wire, this EMT fittings, all of this directly into our procurement system and actually send out a request for quote from one of our partners that we've designated are available for these products immediately into our inventory system.
You can see how this dominoes very quickly in terms of getting everything done very fast.
Hey, Jack. It looks like an extremely comprehensive system that's custom created, fulfills a lot of your business workflows and ties together a lot of disparate data streams and workflows as such. I guess maybe just 30 seconds or a minute overview on how you did this, in terms of the technologies you used, any vibe-coding things that you used or anything like that?
Yeah. Here at Otovo, we largely build all of our own agent infrastructure. We have our own autonomous AI agents that work off of internal and customer requests. If we have an issue on any different system, we have an autonomous system that will pick it up. We call it Nessie, after Nessie the Loch Ness monster. The name is just because it's fun. We obviously use Codex and Anthropic models frequently across the system to build this out. A lot of it is we really love building our own agent loops. When you look at Cursor or any of these other companies, they have one main agent loop design. They said, "Okay, hey, we're going to optimize this for this task. It's going to be for coding.
It's going to be really good at coding." Here at Otovo, we have about 11-18 different ones for each different task. We have a separate version of Iris for emails, for texting, for building out these AI estimates, anything like that. This is mainly for token efficiency and performance. We get very low costs. The month of July has seen us use about $4,500 in total of AI usage costs across the entire company, both Europe and the U.S., on the system. You're going to see that we've really done a very good job at being AI efficient as well, while also building out this system with Codex with Anthropic models and things like that throughout the system.
What was your peak spend? Like you said, $4,500 of run rate. What was your peak spend while creating the system?
In the month of May, we actually had about $16,000 worth of spend.
Okay.
We've now done more with AI with less spend by just tuning in our AI systems. That's where you see a lot of companies now overspending on AI, it's very easy to overspend on AI. Here at Otovo, we make sure that we don't. We're very, very intentional about making sure that we don't overspend. That's not because we shirk away and we just use less. We use more. We just are very, very smart about how we build our agent loops and how we build our agent systems throughout the platform.
What's the average spend that you expect this year to be on a per month basis?
I expect it to be less than $6,000.
Total?
For totals.
Yeah. Staffing on with Jack fully loaded, everything else, total tokens, everything. Anything else that we're spending, which is basically pretty much all of it's ours, right outside Ramp, Stripe, and NetSuite. That's about $1 million a year total. Jack, do you want to go up to the dashboard and just keep it there and just show the different dashboards thing? Give us some time, folks, as we answer questions.
Can you confirm or comment on quality, right? With speed, a lot of times it's messy, there are bugs, there are mistakes. There's just handoffs that are not right between agents or something. How do you guys ensure the quality of the software is as good or maybe even better than what the incumbent solutions were?
Yeah. Here's one of our dashboards that we do actually confirm the ability to do this. You'll see this is our Iris management dashboard, and this is what we're doing to reduce handoffs throughout the system and also manage any errors or anything like that throughout the platform. We have an independent AI rating this system in order to get the star rating effectively up. As you can see, we're really reducing our handoffs even throughout this week as we design the system more and more through this. Of course, no response is not necessarily a bad thing, right? Sometimes people text us a thumbs up or, "Yes, thank you," or something like that. We're starting to get a very significant portion of our emails and texts completed with Iris.
We have 24 of these types of dashboards, as you can see on the main system, that we consistently ensure quality. Each one of these is, we can see the reply, we can see why we didn't reply, the star rating of each and every single one of these systems here. I'm constantly in this system making sure that we can make sure that we have great customer experience.
You can even see we filter out different spam threads and things like that in order to not flood our humans, and we use AI to do that. It's not like we don't have as many false positives on this system. Quite a lot of big companies like us will get quite a lot of unsolicited promotions or different things like that that we otherwise don't want in our system.
Hey, Phil, from the CEO's perspective.
Yeah
My experience has been, you have a big quote tool goes down or big customer service.
Yeah
because some SaaS company decided to change something, and it didn't quite fit with another SaaS company somewhere in the stack, and you're sitting there trying to run around and figure out what's going on.
Yeah.
Here, you just go in, you fix it. The AI Endurance in Nessie is constantly looking for faults and then fixing them. The RSI, and that's Recursive System Improvement that you talked, it's been in theory, we've been doing it since last fall. As recently as 60 days ago, it was basically 5% of the issues, the tickets that are put in or faults within the software were found and solved by Nessie. I think, Jack, that's with the new models, that's approaching 85%. That kind of change in. When you look at the possibilities here, the speed of which you can fix any problems is just, I still am not used to it. It's unbelievable.
Yeah
The quality is really high.
Yeah. Interesting. There's some questions here. First one here is, can you elaborate on your qualifications for leading the AI implementation, Jack? What experience or expertise are you bringing to the role? Or is the approach primarily based on rapidly building AI-powered solutions without a deeper implementation strategy? This guy continues, "The whole solution looks vibe coding." He continues, let's see. "Can Jack talk about his qualifications and why did he get the position?
Yeah, absolutely. I was very early onto AI. I was working with OpenAI on a research preview called the GPT-3 back before ChatGPT was even a thing. I've been working with OpenAI and AI long before anybody else was, back when the system was kind of not really a thing.
What year was that roughly, Jack?
Man, that was 2022. Yeah.
Okay, great.
Began to use that primarily in my work long before anybody else really did, and then even started to use ChatGPT in ways that people really didn't use it at the time, long before anybody else was doing it. Using and developing early versions of Cursor back before it was really a thing that people did. Cursor used to be an email writing assistant, if anyone knew that. Now it's a coding assistant. Honestly, I was just really early. As for experience in general, I have about as much experience with AI coding as anybody else will have, long before anybody else was really into the system. On the saying, "Oh, yeah, it's vibe-coded." Okay. Yeah. We code a lot with AI, but it works.
The only thing that matters on any part of this system is do I get a technician out to that customer? Is that customer happy? That's all that matters. That's all we at Otovo care about, and that's all anyone else should care about. Spending engineering resources and time and money on stuff just to make sure that it is somehow human pure has no bearing on what we actually do here at Otovo. For our motto here is relentless, constant pursuit of absolute customer satisfaction. That's it. We want to fix your thing. That's it. If our software gets it to you the fastest, then that's what we want to do.
Awesome. Can you-
Can I jump just quick?
Yeah, please.
last month when we talked, you had this OpenAI posted on the LinkedIn about Otovo. Can you just tell us what was that about?
Yeah, so-
What are you doing with OpenAI for business, and what's the real partnership under the hood?
Yeah, OpenAI posted us on our LinkedIn, just acknowledging our extreme advancement in using AI at Otovo. The large part of that is really just we're doing stuff that no one else is. Throughout the day-to-day, we are building new systems and Endurance that no one else is really doing. That's really all it is that we are at the absolute forefront because we went all in on this system very early on. I think that's going to be very difficult for other companies to do because there's a large bit of hesitance of, oh, well, what if something goes wrong or whatever. What if this happens? If the quality of the customer's experience is the best in the industry, you're going to win every time. That's what we want to do here at Otovo is you have to take that leap for AI.
I think it's going to be very difficult for existing companies to do so.
Hey, Jack, can you quantify that customer experience for you guys? Where has it been, and where is it now with Endurance operating as the baseline or the newer operating system for the company? Thanks.
Exactly, yes. If a customer calls, texts us or whatever else it is at 1:00 A.M., most service companies don't normally have somebody ready for you. With an AI system and the AI ability to dispatch a technician all the way through, we actually truly have a system where the only human you'll have to talk to is a technician. Why do you care about that? Why do you care that the AI is the only one through? One, obviously our costs can be lower, and what we charge you is much lower. Number two is most service companies do not have someone ready for you at 1:00 A.M. I was in a situation recently where a solar system conked out at 1:00 A.M., didn't have air conditioning, and that's a major problem. With Otovo, you can text us immediately.
We automatically dispatch a technician all the way through the process with our load balancing system, making sure that we send the closest technician, the most available technician to you same day. Someone can be there at 8:00 A.M. before the phone line's even open. That's the real benefit of Otovo. Also, when you are dealing with a sale on anything like this is my system, as you can see here. We have an actual DocuSign integration. Now this link is expired. Actually, no, I can still show it to you. This quote is going all the way through the system, explains to the customer exactly what is going on. We actually got, before we started building out these kind of in plain English explanations with quotes, we were getting issues where customers were like, "I think this is made with AI." It wasn't.
It was made by humans, it was very technical, not understanding. They don't understand what they're buying. This is another thing that we bring to the table. We help customers understand what they're buying. We don't want to fleece anybody. We don't want to do anything like that. We want to just provide genuine, good service. That's why we break these down. Writing up each one of these each time is extremely time-consuming and very expensive.
Doing that with AI now is quite nice to be able to do so.
Yeah.
Also, on a normal one that has not been impaired-
A bigger picture question, do you have an overall NPS score? That might be more of a question for John. Where you can show that trend of improvement. Not only is Endurance as good, it could even be better, given all the examples that you gave us, Jack. It'd be interesting to see it for the entire company. Now, maybe you guys don't have it yet, but that might be something that might be very useful to see in the future.
Yeah. I'll give an answer. Jack, you can chime in. We started looking at our CSAT scores. I think one that would help answer your question, Phil, is that in Europe, where we'd had the legacy, it was quite poor. The service response times measured in months. By the way, that's fairly typical in the industry, whether it's U.S., Europe, even Australia and so forth. It's just a really bad experience. You see that a lot. You write about that quite a bit, about in the media, regulators and attorney generals in the United States and all that. When you look at the U.S., which is just really known Endurance, it's actually well over 4.6 and marching higher. I think it's closer to 4.7, 4.8. Really good customer service and satisfaction.
As these jobs are coming in from our business-to-business manufacturing partners and other partners.
Yeah
We're rapidly burning them through, burning in a good way, getting the problem solved for the customer. It's a dramatically better experience.
Can you quantify this, John? For example, before you guys started with Endurance, what was the average time to service or time to resolution? What is it now?
Yeah, I can answer some of this.
Sure.
With Otovo Cloud, a traditional non-vibe coding, no AI in the platform, no nothing, and HubSpot in Europe, you'd be lucky if we got to you, this is obviously prior to the merger, within the next three to four months.
On average. CSAT NPS was sitting at about 3.3 before we have ripped out the old system.
What is CSAT? Sorry.
Sorry. Just a general metric of customer satisfaction, so customer-
Got it
satisfaction metric.
Out of five.
Yeah. Not good. Very bad.
Yeah.
We had customers that we hadn't contacted for a year, for four months.
Okay.
for emailing us. Actually, when we put in Endurance into the system, replacing the legacy HubSpot system, which HubSpot's number 2 behind Salesforce, a very impressive system in general. We were getting flooded with emails. Iris was getting flooded with emails saying, "You haven't been able to contact us for four months," or a year, or something like this. "Thank you for finally responding. This is awesome." We've been able to track our callback metrics. We've been able to get back to customers. None of this is an easy process. Going from not contacting customers to contacting customers is very difficult, and the backlog is huge. That's what we're doing with Endurance here, is taking these legacy businesses that may have had issues before, and then in three weeks, refreshing them entirely with a brand-new platform that addresses customer concerns.
Hey, Jack. A question on where are you hosting all of this infrastructure, AI agents, what are the hardnesses look like? Are you using any cloud vendors? People are extremely touchy feely when you put AWS in the graveyard, so you putting your hand in a big hornet's nest there.
Yes. I started out building Endurance. I'm a computer engineer and electrical engineer by trade. I have the ability and the knowledge to build out these server systems. I wanted to build it immediately, not necessarily for the cost reductions, which I'm not going to lie, they are very nice. We were spending about $14,000 a month on AWS, I think with the equivalent compute that we have now, it would be closer to $45,000 to $50,000 a month in AWS, especially with their price increases now. We really like to use compute for a lot of different things to speed up the process. We like to own our stack for reliability reasons, really.
We've built out a load balancing system between multiple regions in Europe and the U.S. with a proprietary way of doing Kubernetes and load balancing between those two, along with our own visualization platform and AI hooks on the engineers to make sure that system does not go down as well. We've built out an AWS system internally here at Otovo. The reason why is these AIs are extremely good at the process of getting right down to the nitty-gritty of what the problem is. If you have a layer in between that it can't see the source code of, it can't see what the real problem is, they'll often falter, and it becomes really difficult to solve the underlying problem.
That's why we do everything in-house and on our own server infrastructure, built out in our offices and other locations around the world.
There's a follow-up question here from the audience. This person's asking, "Is that infrastructure included in your $1 million per year spend?" Yes.
Correct.
Yes. Perfect.
We spent a total of $5,000-$7,000 in one-time costs on infrastructure.
That's incredible. A computer alone should be $2,000.
It costs more now.
Yeah
run-ups in hardware and so forth. The point here is that, look, what we've done, and this is directly applicable to any other service company, not just in the power services industry, as I've said repeatedly, this is hard. You got a lot of pushbacks. "You can't do this. You can't do that. You do this. I don't know what this is." Hey, we just went in and did it. Take the technologies, lean into it. You get people that are free from all the clutter of what used to be done. That's old-world thinking. That's gone. You want to have a cost structure. The most important thing, as Jack said, is to deliver the best customer service, period. Full stop. I'll tell you what the other thing, and everybody on this call cares about making money and dropping the cost structure so much.
You have to go in this direction. I don't care what business you're running into, and you're starting to hear that more and more. What we want to do, and I know what Rohit and you, Phil, wanted to show is this is doable. This is being done. What scale is this? We've been doing this for months. Tens of thousands of customers. Tens of thousands.
It's extremely scalable. This is the way the world is going. You've heard a lot about it. Now you get to see it.
This is real.
Yep. Something else I want to emphasize; we have a DocuSign integration directly in our system as well. When we onboard new partners, we get customers to sign the quote. Before, with EnergyAid and other acquisitions that we've had, it was a, "Hey, let me send you the quote. Okay, don't sign that there. I'm going to send you a separate DocuSign account." Oh, by the way, to the company, that costs us $20,000, $40,000 a year on DocuSign. That's at a small scale of one of our small companies, right? It's an overall better customer experience. Oh, you can sign, pay, and everything directly from this one thing. Here, I've signed up here. You attach your payment, bank account, credit card, download the invoice, all right here.
That's the idea, is that we can get a very easy customer experience. We don't have to walk the customer through, "Oh, no, it's in your email. Oh, it may have gotten sent to your spam," or whatever else this is. We can send texts, we can send email directly through here, we can get the link, and all of that is pre-populated in our system here. I also want to emphasize here the customer experience aspect that we were just talking about. This is all of the customer communication that we've had with me. Obviously, I do quite a lot of messing around and testing through the system, but the idea here is you can see everything that is going on with the customer at once. That's for scale, right?
A lot of this is very difficult to aggregate all of this information if you have to go to seven different SaaS softwares, which is exactly the case of EnergyAid. They had to go to a different one for texting, different one for email, in order to schedule this customer. Here, hey, I can look at this customer right here, and I can say, "Oh, hey, wonder why they haven't signed this quote yet. Wonder why they haven't accepted it. Let me see what they were just talking to us about. Let me see what they just inbound called us about," as me as a salesperson. Also see the recording. Obviously, this is a seven-second call, so not exactly anything to put in here. All of this is available directly for our salespeople, directly for our dispatchers, and everything like that here.
Yeah.
Something else I want to show you here on our system is our appointments dashboard. As we're going through Ready to Schedule, our system is continuously updating this with the technicians that we believe are the best on-site. All a dispatcher has to do is take a look, make sure we have the parts, which actually we're going to automate here pretty soon, and then just press confirm on all of these, and boom, the appointment has been scheduled.
Yeah
With drive time accounted for. All of that is not using Google. We built that in-house in our own system. We actually use our technicians' live location tracking, since we have enough density in a lot of different areas, and this is coming at scale, of course, in order to properly move this through the system. We can actually.
We have a bunch of questions here, Jack. Sorry to interrupt.
No.
Just want to give people some more context.
Yeah.
Somebody's asking, how large is the company and how many customers does it have? Trailing revenue last 12 months, call it 50+, $50 million-$60 million. Is that right, John?
I got to be careful here. We're in this red zone, quiet period. We've closed on SunSystem Technology. We're about to close on Green Panel. These are substantial improvements. I would say that the company, the customers that I listed out, that's quite a bit lower than I think where we would be running at right now. Our goal for the year is 60,000. We've disclosed that. We feel pretty good about where that is. I would say that if you look at this is becoming a business that's quite large in terms of anything. We right now believe that we're the largest behind-the-meter power service company in the world.
Great. We only have three minutes left. Just answer this question quickly if you can. How are you guys handling security, managing risks of cyber threats with Endurance?
Very simple on this. We do constant audits of our own system internally. We are able to, because we have full control of our servers and everything in our system, we're able to lock everything down by a lot. We take security measures more than most companies will, just because we have control over the whole systems stack.
Got it. Next one. What position did Jack have at OpenAI, and for how long?
I never worked at OpenAI. I just had early access to GPT-3 and the research preview.
Great. Do you have a relationship with them, like senior people there at all?
We have a relationship with OpenAI, as you can see from the LinkedIn post.
Okay. Next one. Having built the system, what issues will be most challenging for the average enterprise to replicate what is done here?
I think you're going to see a lot of difficulty in handling the volume of data and handling all the little edge cases here. Actually, something I'm showing you here real quick, just as a thing, is we even have our own HR platform that we've built internally, called Pulse, part of the Endurance umbrella. That's because we built it because we needed it, because we didn't know when people were having time off and everything like that for technicians. The hardest difficulty is going to be, A, making the leap to actually doing this, because it's a hard leap. It's not an easy leap.
Yeah.
Also making sure that every single aspect of your company is under something that you can control in order to really start getting the labor savings.
Thank you. Next one. How many employees do you guys have on the system?
About 350.
Okay. How many-
Actually, a little bit more with outside contractors.
Okay. Call it 350, 400. Then how many tech people are working to support this software package?
There are six globally. We have two in Poland, two in Houston, and then one in California, and then me, of course.
Okay. We only have a minute here. What % of your token spend is for open-weight models? Which ones do you use, and what is the framework for how to optimize token spend? 30 seconds.
We have primarily found that OpenAI is actually the cheapest in terms of a actual solution, though we do benchmark open models routinely.
I want to answer Rohit's question, Phil.
Sorry.
What is something that blows me away that we've done, and we can see in the future? We've got one slide says, "Technician's the only human touch point on a service." We've done that. Now, it's not predominant at this point, but we've done that, where a customer has a problem, it's automatically detected, or they text us, then the entire Endurance system orders the part, puts it on the right truck at the right time with the right tech going to the right house at the right time. In all of that, the only human involved is the technician. That's it.
Great. This is a little bit of a provocative question. I think I know what your answer's going to be. This guy is saying, "If the whole pitch is that AI makes it fast and inexpensive to build the product, then what exactly is your competitive advantage? The new investment case seems to revolve around the value of the new platform, yet throughout the presentation you've emphasized how easy and cheap it is to build. If that's true, what's stopping others from replicating it? What makes this platform a defensible asset rather than something competitors can quickly recreate?
This is my favorite question. I love this question.
Go ahead, I'll chime in too. Go ahead, Jack.
To be honest with you, the actual platform itself, zero. That's what we want to emphasize here, is the competitive advantage is the operations, the know-how, the, hey, which statuses do you have in the system? How do you get to a customer? It's still the people behind it and understanding how to actually execute in this service business. We've been struggling with this system throughout this whole past year. This is not an easy cut-over. It takes courage as a company to be able to do this. It takes an absolute decision-making down from the top to say, "Hey, we are doing this." That is going to be impossible for 99% of companies, to be honest.
Yeah
through this process. I personally believe I could publish the source code here for the next day, and you can all have it, and the value of the company would not go down at all. It's the value of progress. That's what I want to emphasize for Otovo.
Great.
Yeah, I totally agree. This is the big problem with AI, and the challenge for all these CEOs, I really sympathize is, this is incredibly difficult because you got to reorder your entire organization. You have to think about is tasks, not roles. Think about what AI can and cannot do. Like Jack said, you have to have the courage. We've seen some of that pushback in some of these right here. By the way, easy was not ever used in this phrase. That was something that that gentleman or lady put in that question. That was not a word that we've used. This is hard. Is it cheap? Is it very inexpensive? Is it a competitive threat to every single existing service company, not just SaaS? Because instead of software as a service, it's service as software. It's collapsed down.
You have to have all this together and rethink the entire organization. We've got a huge head start, and along with the companies that we're acquiring, we've got a huge head start on the roll-up. We like our competitive advantage very, very well.
Great. Rohit?
Just one last one-word answer, please. Knowing what you know about OpenAI and Anthropic, which company IPO stock would you buy?
Valuation excluded, of course, Rohit.
Any answers? Jack?
Jack, go ahead. It's your question.
I personally would do OpenAI.
Okay. I got you. Cool. John, you want to chicken out, or do you want to say something?
No, thank you for the opportunity. I think what you've seen here is an exciting preview of the future. What has been talked about often is, can this be done? Can you develop your own enterprise-wide software platform? How do you embed that into your organization? How do you take agents and displace human tasks? How do you drop down your cost? How do you improve your levels of customer service? What we've done is we've done that. You'll see it in two ways. One, customer service and satisfaction improvement globally. Second, financially. When you look at what Otovo has accomplished and what we will accomplish, the speed is breathtaking. Two weeks is a long time. You have to be young to keep up with this and just keep up with the hours, I would say. I really mean that.
When you look at what the future holds, it's exciting. There's so much change that needs to take place. You just have to have the courage as investors and management to take it on. You know what? There's no other way to move forward in life. You got to embrace the change. You got to embrace AI.
Great. Well, let's wrap it up here. Thank you all for attending. Otovo team, Rohit, stay with me. Thanks, guys.