Robotics In The Workforce – Andrew Ashur, Lucid Bots

Robotics In The Workforce – Andrew Ashur, Lucid Bots | Full Episode

William Bissett: Um, this is our monthly speaker series. We call it Charting Opportunities. It is geared towards small business owners, um, and helping all of us navigate the difficult things that come our way. Um, we’ve had this Andrew’s our seventh speaker. We’re gonna take two months off because nobody’s here in June and July ’cause it’s hot and humid. Um, where, but we’ll have a podcast or two to kind of supplement it. Um, but as our seventh speaker and our last one before the summer break, we, um, are very fortunate to have Andrew Asher. Andrew is the founder of Lucid what was originally Lucid Drones. It’s now Lucid Bots, potentially changing names again, it seems like with email address.

Um, but anyways, Andrew was actually on my podcast, what, three years ago? Yeah. Um. And that a podcast geared towards startups called Speaking of Startups or the Charlotte Angel Connection. And so Andrew has done a phenomenal job of growing the business from a college student, um, to now married with a child and 65 plus employees. Yep. Um, you’ve raised a recent round, um, so raised capital to continue to expand and grow, um, off of robotics and AI. Um, so he knows the topic in and out. I told him earlier that won’t set expectations very high because I know nobody else like him. Um, he’s the most informed AI and robotic person I know in town, um, and also the only one.

So, um, with that being said, Andrew, thanks so much for coming out. I’m gonna give you the floor. I’ll come back up in a little while. Fantastic. Awesome. Thank you.

Andrew Ashur: All righty. Well, thanks for having me. I, I’m really grateful for the chance to talk more about robots and AI. Uh, it’s probably a topic you’re reading more and more about. I’m gonna spend a few minutes going through slides just to give a general overview about the topic and a little bit about what our company does, and then William’s gonna come back up and we’re gonna dive into a deeper conversation. I think what often happens, whenever you hear the word robots is there’s this initial fear of are robots just taking jobs? And from the start, we’ve always asked the question of. How can we use robots as a mechanism to uplift where the point of robots isn’t to replace a human touch, it’s really to extend human reach. And I’ll talk a little bit more about our story and what that means. Um, what I really wanna start by talking about is the why and the problem. We’re, we’re running outta workers in a lot of these critical jobs that need to be done. If you look at something like the construction industry within the next decade, over half the workforce is set to retire, and the cleaning industry, which, which is our primary focus today, you’ve got about a 200% annual turn turnover rate. So imagine if you have a hundred cleaning techs, you’re hiring 200 a year just to stay steady with your current revenue projections, and of course, declining in birth rates on top of all of it. So these are massive markets that are desperate for a lifeline and we gladly toss them on.

One of the fundamental problems that we see is AI is trapped behind a screen. Today when we kind of zoom out and we think about the venture landscape in general, we. We’ve innovated across bookings, customer success payments, all these things that touch the services industry, but we haven’t solved the most critical problem, which is people and just getting the actual physical work done. And one of the things we realize is with a robotic body, you actually allow AI and software to break free from that confine of a screen and do physically productive tasks. So when we just think broadly about it, this is the same story with a slightly different application. If you look at the software and SaaS world, anytime you had an ancient tool that relied on a pen and paper or a spreadsheet, and you were able to go in and write software, you could meaningfully automate it and create really good ROI For companies in the world of AI, the data you often need here as you look at something like a large language model.

Is text-based data. And the great news is there is we have an abundance of written word that can be found readily accessible online. We’re now solving a similar problem, which is ancient tools, but they’re the physical kind, they’re the squeegees, the pressure washing rods, you name it. But for AI to truly work there, you need a different kind of data. You need this spatial data. You need the robot sensor and motor data. Uh, and that’s a much different problem to solve for. So let’s see if it will go to the next slide. There we go. The three themes you probably keep seeing surfacing in the news are robots manufacturing the United States and physical AI. And part of why is because there are, this is a multi-trillion dollar market. This is one of the biggest problems and opportunities we face as humanity. A thing we often like to joke about at our company is that the, can you go ahead and hit the next slide? Thank you. Uh, the worst time to think about the current thing is the current year.

Um, many years ago, in the fall of 2017, I had this crazy idea, uh, and it happened a little bit by human accident. While I was studying the highly relevant fields of economics in Spanish at Davidson College, I saw these window washers on a swing stage, and it was a windy day. The platform they were on got caught in the wind and just started slamming against the building. And you saw the workers up there, white knocking the rails, hanging on for dear life. And as this was unfolding, I asked what seemed like an obvious question of how could we leverage technology to make this job safer, to make it more efficient? And we’ve quite literally been through the school of hard knocks. Uh, the first year out, I ran my own cleaning company to learn the pain points of our customers. I did it by hand. I did it by robot. I’ve got, uh, chemical scars to tell the tales of what it looks like when it goes wrong by hand. And one thing I learned along the way is I used to think I was brave until a customer invited me out on a 30 story building, 28 stories up on a swing stage.[00:06:00] I took one step out and I mean, I’m, I’m like ballet stepping. I’m trying to be as gentle as can be, and you immediately feel that platform start rocking. So, uh, we’ve been building again for years and years. Last year we were honored to be recognized as the fastest growing robotics company in the country. Uh, at the same time, it really does still feel like we’re just getting started. When we look at the overall opportunity and problem we want to solve. Uh, internally at Lucid Bots, there are three fundamental truths about the future. We often talk about, the first is that robots are the future of these traditionally dull, dirty, dangerous jobs. The kind of jobs that are harder and harder to find people to work, yet the work still needs to be done. The second belief we hold is these robots need to be built domestically. As you look at our economic resilience, National security data privacy, relying on robots built outside of this country poses a huge vulnerability. Then the last belief we have is every single thing that moves will be automated, whether it’s your car, a robot, or ev, anything in between. And within the the broad ecosystem of technology, we believe this combination of automation and robots is what unlocks a world of human abundance. Like really nothing else can.

So this is a, a very short video just to give you a, a glimpse of what our core product the Sherpa does today. Uh, when we first started, we thought about the obvious use cases like hotel windows, office campuses, but we’ve had customers clean anything from industrial storage tanks to stadium domes, hate speech off the side of bridges. You name the elevated structure, and on average it’s about two times faster. In some cases, it can be as much as eight times faster. So let’s imagine for a second that you’re running your own exterior cleaning company and you got hired to clean a 10 story building in town. Usually that would take AC crew workers days to weeks to get the job done, and our robots can do it in a matter of hours. You’d have to rent, give, or take 120 foot lift, which could cost over $10,000 a month to rent. Our robot costs a fraction of of that. And last but not least, you’ve gotta elevate your workers on these lifts, ladders, scaffolds, putting them in harm’s way. Unfortunately, this equation of humans versus gravity, it’s not one where we win, but with our technology, we can safely relocate workers to the ground. So you no longer have to read about these headlines like the one at the bottom. These are just a a few examples of the types of surfaces our customers clean. Anything from windows to these. Slate roofs, industrial storage tanks, and again, the, the list goes on and on. It cleans just as well, if not better than you can by hand and produces those transformative efficiency and safety gains.

So, I won’t spend too much time on this slide, but just to speak to it, uh, in one example, we had a customer who was making about $30,000 a month per cleaning crew almost double their top line revenue and add $300,000 over the course of a year. That’s before you factor in things like insurance, saving lift savings for rentals, landscaping savings, and maybe my favorite example. Um, this was Occasion Dome down in Louisiana. They spent over a million dollars to seal the roof to prevent future mold growth. How do we think that looks? Pretty bad, right? Uh, what they didn’t account for was the road fumes. There’s a highway nearby and we had a customer then go in with one of our drones and they were able to clean it in about two and a half days and make really good money while doing it. So for us, our product strategy at the company, uh, can you go back? One slide is really split into three distinct things. Although today it looks like we’re building a cleaning drone and a pressure washing robot, we talk about the strategy of brain frames and payloads, where the goal is to make one universal AI robot brain that we can apply across any type of robot platform. Today we focused on an industrial drone and a ground-based robot, and then we attach these different tools or payloads to this modular frame. So at the start, we’ve been very focused on cleaning, but over the years we’ve done anything from. Uh, construction ceiling applications to agriculture, to painting delivery. The list goes on and on. Uh, specifically as it relates to AI, I’m sure everybody in here has used chat, GPT or other large language model interfaces. We are now starting to see more and more VMs, vision language models come online, but there’s still this gap for robot data and people are starting to work toward what are called these robotic foundation models. And you can collect data in a lot of different ways. You can do it through a simulator, you can try to find some online data sets, but they’re pretty limited. Uh, or you can actually build robots that have real world utility and collect data out there in the wild. So from a data perspective, data really is gold when you look at AI, but most robots today are built for warehouse environments. While that’s good and there’s a lot of value to be had, they’re stationary where usually the work’s coming to them, sometimes they’re quite literally bolted to the floor, and it’s a pretty limited amount of data you can collect. And we’ve really focused on building these industrial applied robotics that go into dynamic environments. Instead of the work coming to our robots, our robots go to work where they’re on a different job site each and every day, which allows us to collect a pretty diverse array of data. And with that data, there’s a lot of different value we can provide customers. One is on the predictive and proactive side of the robot maintenance, the same way I could go into the. To the, uh, parking lot, measure the tread on your tires and know when should you replace your tires. We can actually look at our data and understand when does a battery need to be replaced? Does your motor need to be replaced soon? And so on. Beyond that, though, we’ve also got the ability to collect visual data while our robots are operating, so we can train our robots to look for things like when windows need to be recalled. When there are cracks in buildings that might warn something like a repair. And then of course there’s the holy grail for us, which is making the robots increasingly intelligent and autonomous. Um, before we talk about the glory of all things artificially intelligent robots. It’s really important to address that. Scaling a robotics company is unbelievably challenging. First, you’ve got this technical, um, coalescence of mechanical engineering, electrical engineering, and software engineering that all have to play nicely together. That’s a hard enough problem in and of itself. But then you start thinking about standing up a production line, QA, QC, how are you gonna scalably support these customers? What is your repair department gonna look like? You’ve now been building robots for years, so how are you going to control both software versioning and hardware versioning, and then everybody’s favorite term these days, supply chain? How are you gonna manage supply chain across these versions, especially when you’re a company that likes to move quickly with product?

To, uh, maybe go ahead and give a little glimpse into where we go from here. What we’re really working on is taking a chat, GPT, like interface and putting that on the brain of our robots where you can truly make them collaborative. Today, when you think about automating robots, you typically need a very advanced degree and have been in this space for years and years. So it’s, it’s highly technical. Um, what we wanna do is democratize access to robots, both in terms of use and programming where anybody that’s capable of speaking English can use their natural language and communicate with the robots the same way they would a coworker. So they could say, Hey, Sherpa, uh, up there, five stories, two windows from the left. I see a bird just pooped on the window. I want you to go and clean that specific spot. And without writing a line of code, the robot will intuitively understand the work to be done. And then based on the success of that job, learn from it the same way a normal person would learn from it. There’s a lot more to dive into, but I just wanted to go ahead and set the stage for robots and all things AI and uh, invite William back up here for a discussion.

William Bissett: I think it’s great if we had a, um, a first of all interface. Um, second of all, I think it’s great that, uh, we have a robotics AI person up here talking and we are manually back there changing the size. I was already remote control wasn’t working. Um, so anyways, sometimes technology doesn’t work when you want it to. Right. Andrew? Uh, unfortunately. Um, which I don’t know, every time I hear about AI, the excitement in Negros. Mm-hmm. And then the fear level in Negros as well. Right. Um, like where do we end up in terms of a workforce, if the machine. Whether it’s a physical machine Yeah. Washing windows or whether or not it’s an actual like computer with, uh, chat GPT or whatever it is. Yeah. Is doing that work. Where does, where does the workforce end up in 10, 20, 50 years?

Andrew Ashur: Andrew? It’s a, it’s a great question, and I often like to go back to history in order to understand where we’re going and. I just rewind the clock. A hundred years if you’re in the US workforce, a little over 40% worked in agriculture and farming. Today that number is about 1.6%. Yet if you look at hunger levels, they’ve never been lower. So technology usually creates new opportunities. And I remember, um, and there’s this old econ lesson about the the 10 fisher person island. And the theory goes, 10 people get trapped on an island. And at the start, all of ’em have to go out there and they’ve gotta take these spears and. Try to get the fish just to be able to survive. And then one day, two of them have this brilliant idea, what if we weave together a net? And it worked and all of a sudden two of them could go out and collect enough fish for that society. So one of them went on to focus on building housing and another went on to focus on farming practices. So what we see is the, the world will continue to evolve and new jobs will be created. What those will be, I think it’s still very uncertain and, and the one thing I spend a lot of time thinking about is just the rate of transformation and change. Uh, with AI, if you look at the innovation cycle, it’s going faster and faster and faster. And part of what I worry about is like, what is our country’s infrastructure to adapt to that change? Because if you study science, what you’ll learn is in order for an organism to survive, its rate of learning has to be faster in the rate of change in the world around it. And if you look at technology, it’s changing really, really fast. And the conversation I’m having with a lot of business leaders right now is. How are we setting up our workforces, our people to embrace that change?

William Bissett: Uh, so many questions that feed off of that, right? Yeah. Um, so, um, but we’ll ignore. Does the, uh, does the college infrastructure and everything else support. Mm-hmm. The next generation of kids, we’ll stay away from that. Talk about, um, the introduction of robots into the workforce, right? Yeah. Like what, um, obviously opportunities evolve from it, but, um, what are some of the issues that y’all think about or that businesses are coming to you and thinking about in terms of, well, if I do this, what does it mean for the worker alongside of ’em? Mm-hmm. How do we control, right? Like what are some of the ethical or moral issues about, um, yeah. Introducing robots into, you know, side by side with people.

Andrew Ashur: I, I think one of the ones we talk to customers a lot about is accessibility. Will the workers I currently have on these job sites be able to learn and embrace this technology? That’s something we really focused on from the start. I, I was the cleaning worker and I’m like, I’ve gotta figure out how to use this. And again, I, I think part of the beauty of. Uh, having, uh, an atypical background in liberal arts and starting, let’s call it more or less a engineering and robotics company, is I had to think about how do I make this simple? How do I make this something that anybody can access? And one of my early tests was, uh, teaching my grandma how to fly our cleaning drone, and she picked it up in lesson than five minutes. Uh, and this goes back to our central belief, that the point of the robot is to extend reach versus replace touch. But that’s not the case with all the robots that are being built today. Some will work, but it takes a tremendous amount of training to the point you’ve gotta bring in a different type of worker versus taking the person that’s already working that job and allowing them to two x or three x their output. So we have a lot of conversations with business owners about that, but one of the interesting things that surfaced in those conversations is. The desire for the next generation. Uh, I remember early on, um, we were talking to a national cleaning company and they’re like, well, William’s been cleaning for me for three decades. He does a great job. And, you know, no, no robot’s ever gonna clean better than him. And I, William was at the table in this instance and I said, you know, what do you, what do you, do you have any kids? And the answer is, yes, I’ve got a few sons. It’s like. Well, what do you, what do your sons want, want to do? What do you want them to do? And he immediately started saying, well, not window. And he, he stopped. I was like, wait, finish that thought. Uh, and what he said is, I want something more for them. So even the person doing the work, he was wishing that his kids would have something even more fulfilling. And I think as you abstract what it means to be human, it’s not the ability to do a, a physically repetitive task. It’s this ability to be creative, to problem solve, to build community. So we see robots as a way by which we can empower people to solve increasingly meaningful problems. Again, going back to the agriculture example, so I I, I could be pretty long-winded, but those are a few of the thoughts that surface.

William Bissett: Yeah, no, absolutely. So, uh, important to note too. So when Andrew says, um, he built it, you started off using off the shelf drones. Yes. And they didn’t work as he needed them to work. Um, so Andrew, the economics and, um, major with this minor in Spanish did the only thing that he knew how to do, which was to build the drone himself. Yes. Um, and then the software didn’t work the way you wanted it too. Mm-hmm. So he did the only thing that a economics major and business minor would do, which was he built the software, um, to power the drone as well. So when you talk about building it, you, you literally built it.

Andrew Ashur: Yes. Literally. And the good news for everyone in the room and our customers now is. Our engineers, we, we’ve gone ahead and we’ve calculated this. On average, they’re about 88 times smarter than me, so I was just a catalyst. And then it’s been, uh, hiring great people ever since.

William Bissett: Um, so, um, so you built the drone for window washing. You tell the story about that. Right. Um, the next one that you’ve introduced is the, is the, um, the pressure washing robot, right? Yeah. Just talk about the evolution of that and how it came about. And then how, how somebody like you walks around and sees, ’cause you see things differently, right? You’re thinking about constantly solving workforce issues. Yeah. How do you think about the next couple years?

Andrew Ashur: Oh, that’s a great question. So the first one is the bot and then the next five years. So, uh, specifically to the LA robot, the pressure washing robot, um, for us, we’re very much, everybody says they’re customer centric, not everybody is. Um, so take, take that with a grain of salt, I promise. We are. And our customers would tell you that. But it was through, uh, just a lot of customer demand that we decided to build it. I think when you first start a company there, there’s always that question of like, when will we expand to offer another product? And the cleaning drill market was huge. So it’s like, why? Why start with that? But over the years, we would ask our customers for feedback. The same thing kept surfacing, which is, Hey, I’m out there cleaning these vertical surfaces. There’s flat work I want to do too. I don’t wanna do it by hand. Could the drone clean a flat surface? Uh, and again with, with my very, uh, relevant econ major, I did the physics calculation and I’m like, you know, flying an object and surface on ground, that’s not the most efficient power equation. Uh, so we just kind of kept listening, but didn’t really build it till we had a customer come in. And he’s like, I need a pressure washing robot. And I was like, well, explain it to me. And he just held up his hand. And at first I was like, are you going for a high five? What are you looking for? I, and he’s like, no, look at the bones in my hand. And he’d been pressure washing for about a decade and the bones in his hand had started to fuse together. And at the time, his daughter is about nine. He’s like, it hurts to hold my daughter’s hand, and that is why I want a pressure washing robot. Because he’s like, I’ve been doing this my whole career and I’m starting to wonder if I need to find a new profession because of how much of a toll this is taking on me. So that’s how we started building the, the ground-based pressure washing robot. And I think when we walk around, the first thing we see is like all the opportunities for our existing robots. But the next thing we do is we just, we listen to the market. Like what are people asking for? What are the biggest pain points customers are having? Uh, and the most recent one is we’ve had a lot of construction companies reaching out to us. They’re like, Hey, wet, curing concrete, applying these ceilings to tilt up wall applications. It’s hard to find people to do those so. We’re in the very nascent phases of exploring those alongside some early customers, but it’s a, it’s high in demand.

William Bissett: Yeah. So you were out at CES earlier this, uh, mm-hmm. Year back in January. Right. So, uh, which is full of new technology, new robots. What did you see out there that was interesting, um, that caused new questions or concerns? You spoke on a couple different panels. Um, so what’s going on in the world of robots and AI that you’re paying attention to that No, none, none of us are seeing.

Andrew Ashur: I, I think one of the big things is always to look at where’s the money flowing. And for the first time ever, you’re starting to see more and more money flow into hardware and robots. Uh, at least on a, a more. Significant level than before. And part of that is, uh, software is becoming increasingly easy to build. And I think if you actually flip one slide, I might put this in the appendix. Let’s see, uh, one more, uh, no to the forward. Sorry. Keep going. So our robot’s back there. Ah, there we go. Um, so this was, uh, on the left was an example of a Python script I wrote back in 2018 to detect the color orange. ’cause one of the, the working thesis I had was, Hey, every construction worker has orange cones in the back of their car. Maybe we could do some localization and automation based off of this. It was so hard, it took forever to get right and it didn’t even really work that well. Uh, and last year, uh, when we took this screenshot, you could literally type in keywords and it will automatically detect. These different things it sees on a screen. So it’s becoming like, again, the rate of innovation is accelerating at an increasing pace. So at CES we saw more and more people interested in the hardware, in the robots themselves. And I think one of the, the things that stood out to me is how many humanoids there are. Yeah. Or people talking about humanoids. Uh, and then I was on a panel where I like to joke, I think the collective market cap was like 150 billion. Uh, Lucid Bots was on that. We were not a huge contributor to the. That market cap calculation. Uh, but it was like people from DHL and whatnot, uh, and they were actually pretty bearish on humanoids themselves. And a lot of it just comes back to thinking simply, uh, the human body. Uh, it can do a lot of different things, but often in inefficient ways that still require tools. Um. So it, it’s, it’s fascinating to see people investing in what looks common, which is the human form factor. But if you’re talking to the people in the industry, the robots that are actually being used, they look nothing like a human. And probably my favorite example was a humanoid company showing the video of this mass production humanoid facility they were producing. And it was just to build the mask. In this one video, they probably had a dozen stationary robots. Building a superficial mask for the humanoid. So if you’re wondering like what robots are gonna win in the long run, I don’t think they’re necessarily gonna look a lot like us.

William Bissett: So the stationary component, the old school, uh, Ford factory where it’s still automation static type thing is where you’re seeing.

Andrew Ashur: Yeah, it’s where the high. Adoption has been to date, you’re starting to see more and more folks like us that are thinking beyond just the warehouse of what are the drones and robots we can build to do these physically productive tasks on say, a job site.

William Bissett: Yeah. Um, so where do you see, um, do you see continued expansion? Like I, I. Keep thinking about like the Ford factory, right? Yeah. You think about Tesla and what Elon Musk is doing, um, and automating away all of those old school manufacturing line based positions. Mm-hmm. Is that where you’re seeing a lot still of money flow in where we’re completely automating that process and then even in that world, do we still need that human contact in those forms?

Andrew Ashur: You, you’ll always need a human in the loop. To some extent, um, but with robots, the number of humans in the loop to produce the same output will go down. Uh, there’s certainly a lot of investment going into these industrial warehouse based robots, and that’s badly needed is we look at manufacturing in the United States. You ask, well, how did we end up in this position where so much our manufacturing got outsourced and it was a, a labor arbitrage? But when you think about the end state of manufacturing, it’s automated. So all of a sudden the efficiency gains and equation starts to change, and all of a sudden it’s like, well, who’s got the cheaper energy source? Uh, so that’s a whole nother, uh, topic. But yeah, you’re seeing a lot more in the warehouse, but you’re also seeing people starting to think about, um, jobs beyond just that. So you look at the largest industries like cleaning, like painting. We’re not the only ones thinking about these types of robots.

William Bissett: So you’re thinking long term painting robots?

Andrew Ashur: Um, I mean, you can paint the outside of a building with a flying drone. Um, as long as you can control your spray. Yeah. We’re we’re thinking industrial settings to start that are monochromatic and maybe more forgiving than, uh, a residential home with somebody who really doesn’t want paint on their windows.

William Bissett: Well, whatever. Yeah. Um, well come over to your house to buy the first one. Yeah. That’s awesome. Yeah. But no, you showed a picture earlier. I had a friend of mine whose family owned a, um, water, uh. Water tower cleaning business. Okay. Back in the nineties. Yeah. Um, and you’re taking that business and you’re displaced, not displacing, um, uh, but the workers that they had, you can do twice as much. Right. And plus, they don’t have to be in it as long. I mean, it’s completely, um, disrupts that industry and a ton of others like it, uh, on that.

Andrew Ashur: One specific picture. Um, before they used our technology, it was a two week job with four people on a lift, and then that job for the drone was five hours and 43 minutes of flight time. So it was just a, a night and day difference for them. But when the job was being completed, it was August in Louisiana and it was like 110 degree heat spell. The pilot did a lot of the flights from the air conditioning of the truck. So a different way to get the work done too.

William Bissett: Yeah, no, absolutely. So, um, I’ve got a thousand questions I’ll ask Andrew. Um, but I, I mean, different space. I just wanted to see if anybody had anything specifically to any industry whatsoever, um, or questions or concerns to you. Yeah. I need to answer Brian. Thanks for, for joining us today. We have a client that’s in the. He has services that support the construction industry. Okay. So he does dust control. Mm-hmm. He does street clean. He does. He empties dumpsters. Right. He does power washing. Is the robot at a place and the X where saw this screen where you could deploy it and say, okay, this street needs to be cleaned now. Yeah. Now the dumpster can monitor by tonnage. Right. But it might not be tonnage. Mm-hmm. Or you could look at things or we had a. Place now where you could have a robot and replace the project manager or the human right that could fly to its 50 sites in greater Phoenix. Right. And tell them that. Mm-hmm. That place now, uh, yeah.

Andrew Ashur: So one of the things we built out is a fleet management tool. And what we’ve heard from customers is, uh, like a great example is gas stations. They need to be cleaned regularly. It’s often at two in the morning, and you as the, the regional manager, the business owner. You’re not out there at two in the morning checking that all five Circle K’s within 20 miles got cleaned. Um, but you can actually look at the data to prove like, yes, the robot was in this spot. Uh, and now we’re starting to layer on some, some visual additions. You can actually see like the pictures or videos of that work being performed.

William Bissett: So in that, uh, follow up question. Sure, sure. So in that case, does he buy the robot? It’s a tactical question, right? Yeah. Does he buy the robot and you teach him how to use it? For these services that you just expect? Yeah. ’cause that would be every time we want to do something. You’re very expensive or Yeah. You buy the robot and kind of explain and teach.

Andrew Ashur: Yeah. We’re much more the technology provider in this ecosystem and for us it’s, Hey, here’s the shiniest tool in your toolkit, right? It’ll help you do more jobs and less time with less liability. So most business owners look at this and they’re like, oh, like more revenue, better profits per job, lower business risk. Uh, and then we built out online training, and we’ve got an in-person option as well.

William Bissett: So when you think of that and you say that out loud, you instantaneously think that you just replaced the bathroom attendant, but you just created probably two jobs behind the scenes in order to manage that as well, right? Yeah. Um, you need somebody to help understand what it’s doing, train the system. Um, and then create the training system in order to educate the people on the ground. Mm-hmm. So you’ve got multiple to, so to your point, um. One job’s being eliminated, but other jobs are being created to manage it well.

Andrew Ashur: And I think another important thing is like there’s this whole job and workforce that’s going to grow out of servicing the robots themselves. They need maintenance. Has anybody ever made it two years without getting their vehicle serviced? Because if you do, I wanna know what car you bought and I’m going to buy it. Yeah. Uh, the same is true for robots, but for some reason a lot of people have this perception of, oh, I buy a robot and there’s never going to be an issue. It is so far from the case, ’cause again, these are on these, these dirty, dusty work sites. My clicker for an example, in case anybody was wondering, it’s the early stage robot. Yep.

William Bissett: Um, so, um, any Mary few questions? Sure. Um, you take the window cleaning. Mm-hmm. Are you disintermediating that mold, meaning that you are selling to the property managers to clean the windows, or are you selling to the current window washing companies to make anything more efficient? And then the next question is, how long is it time to go to market? Yeah. You have about a dozen categories up there. You are only in two. Mm-hmm. If you wanted to add a third, what’s the go to market? Yeah. Start to finish.

Andrew Ashur: Of course. Uh, these are great questions on the first one. We, we very aim to serve the market as it exists today. Our primary customer tends to be the cleaning company, and that’ll be anyone from like an owner operator that’s got one or two people all the way up to these national accounts that have, you know, multiple service trucks in each of the major markets. At the same time, there’s this other segment of the industry of facilities management where people will have. Uh, a facility or a group of facilities and a team to manage those. So, like universities are a great example. So we work with several universities, Richmond, a bunch of the UC schools. Um, so we will also sell directly to some facilities management people as well. And then on the, uh, the go to market, I’ll tell you, it’s variable for the cleaning drone. Uh, again, we spent the first year doing the cleaning service to really, uh, nail it and make sure we were able to build the right thing. And then pretty quickly we found there was a lot of market pull. I, I often think about product market fit as like a river and the faster the river’s flowing, the more people are pulling you in. And even like today, about 95% of our deals are inbound. When a lot of people see what we’re doing or they see the competitor down the street with our technology, they’re like, yeah, that makes sense. It’s, it’s a pretty straightforward product. Um, with something like the Lavo bot, it’ll now be, uh, let, let’s call it like 18 months from when we had that first thought of, yes, we should invest time and money in building this. To getting it closer to a state of production. And sometimes it goes and, and let’s call it like ebbs and flows. So we went early to market with the Lavo bot, with a manual version RC, and what we learned is it was incrementally beneficial. And the value props were, hey, it doesn’t get tired over the course of an eight hour day. So you get like, on average 25 to 30% daily efficiency gains. But the business owners, like they, they didn’t really care. They’re like, that’s not transformative, William. He can go out there and push a surface cleaner, and I don’t care if he gets tired, if he handed it to William, he didn’t want to hand the controller back. But we, we got to this point where we sold our first several dozen versions of the manual one, and we’re like. We’re putting way too much effort into selling these, and what the market kept asking for is, Hey, as soon as this is autonomous, even if it’s light autonomy, it’s a no-brainer from an economic perspective. So for us, we’re like, well, we’re investing in supply chain and training. Let’s go ahead and hit the pause on this until we get it autonomous. And then we still have our, our core product of the cleaning drone to support us from a revenue standpoint.

William Bissett: Super helpful. Thank you. Yeah, so you can think about that, uh, instantaneously when you, um, sell into the existing window cleaning people. Like their insurance cost. Having those people up on the, um, on those lifts is astronomical. Right? Yeah. And as, as soon as you put that person on the ground, that insurance cost and you still got it ’cause you got this machine flying up in the air. Mm-hmm. But it’s gotta fall through the floor. You go from two people on those lifts to one person on the ground, you can just see the profitability for the, and then eventually you’re gonna have price erosion. Mm-hmm. As the market realizes they don’t have to pay as much for it anymore. Yeah. So it, um, all helpful. Um, we’ve got probably seven minutes left. One question I had was, so we’ve got construction, we’ve got manufacturing, um, we’ve got a couple different kind of classic industries where hard manual labor, right? Mm-hmm. Um, when you’re talking to people and you’re doing the, the early stage groundwork, what are some of the things that they should be thinking about? Um, or concerned about, or, um, optimistic about, yeah. When they say, I want to try to put something into my workforce to make it more efficient, to eliminate the burden mm-hmm. From my workers. Like, what are the positives and negative things that should be on their mind Yeah. As they reach out to somebody like you or just start brainstorming about how things might develop.

Andrew Ashur: I think that the, the first thing is like, what are the questions to ask? And depending on what industry you’re in. There are typically three questions we urge customers to think about. The first is, five years from now, do you think there’s gonna be more or less robots in your industry? Uh, the second is, five years from now, do you think the robots are gonna be more intelligent and capable at the job than they are today?[00:37:00] Then the third question is, would you rather be on the forefront or backend of the adoption curve with that technology? Um, it, it’s an inevitability in a lot of these industries. I, I tell you to do a lot of vetting and diligence on any robotic provider you work with. Uh, the earlier the company, obviously the more growing pains they’re going to have. Sometimes it’s still great ’cause you’re first to market and you get to learn alongside them and you kind of get the white glove treatment and you’ll win more customers as a result. Uh, the one thing I’ve seen, whether it’s our customers or like other founder friends running robotics companies. Robots are a profoundly good sales and marketing tool. Um, a lot of our customers, they’ll jokingly say, you know, we’re actually not saving anything on labor ’cause we have to hire somebody for crowd support. ’cause typically if you’re just cleaning a building, everybody keeps on walking by. All of a sudden you throw up a drone or put a robot on the ground, it practically becomes a spectator sport. So it’s been really good for getting more jobs for customers as well. ’cause people will like to see, hey, you’re leaning into innovation and pressure washing. It sounds really simple, but if you have a knucklehead out there who’s not paying attention and they go too slowly and they etch concrete, or they go too quickly and they leave these dirty streaks. And you’re working with a high profile client, like a, like a Circle K, like the Panther Stadium, you can do an unbelievable amount of damage, not just to the reputation of the company, but also the facility itself. So people also like to think about like, what is the robotic consistency I’m introducing into the work by leaning into a robot to do the the core function.

William Bissett: Ah. Carol, Gerald, you know, I was thinking about in light of the administration and the, the effort to bring manufacturing back, I haven’t heard a lot of discussion about AI and, and robotics. Um, and it seems like there should be more of an interface between those two. What do, what do you think?

Andrew Ashur: Uh, I would agree. I, I think like right now, what we’re trying to build at our company is essentially this flywheel of AI and robot. Where we want to give our robots the ability to perceive, reason, act, and learn essentially on their own with safety guardrails. Um, but today, you know, you’ve gotta have a person that sits down and makes those improvements, that runs things through a simulator that does the debugging. But oftentimes you can just take whatever the, the debug code is, put it into something like chat, GPT, and it can solve it on one shot. So we’re trying to create this flywheel where the robot can actually do that work itself and become, uh, let’s call it an AI agent on its own. So I think it’s a combination of the two is we’ve gotta build the physical products and supply chain back. ’cause that that’s truly what creates the surface area of opportunity in the body for AI to then be deployed for these different use cases. But if you build great software and AI and you don’t have the hardware and dependable systems to compliment it, it doesn’t do a whole lot of good.

William Bissett: No. Um, one more question you said earlier, if you’re thinking about, and you, we all think robots will be more prevalent, we all think robots will be smarter. Um, and then to be on the forefront of it, uh, do your proper due diligence in terms of who you’re working with. Mm-hmm. Um, and I think about how to do proper due diligence. Yeah. Right. So, uh, somebody that’s in it that’s grown a company, how do you do proper due diligence on somebody that’s on the bleeding edge of something like this?

Andrew Ashur: I, I think, again, most robots today, they’re going to be expensive. They’re not like a, you know, $200 purchase. Uh, I do a site visit, like go out, meet the team, understand who are the people behind the technology. Um, second, I would ask a lot of questions about their support. Like, what does warranty look like? When do parts start to wear? What is your repair plan? What are the different tiers I can get? And if they kind of shrug their shoulders and go, you’re not gonna have issues, that probably means they don’t have a lot of field data and people using their products yet. Uh, and then I would ask for as much customer information you can get on people who’ve actually incorporated it and what the ROI has been, because so many robotics companies, they’ve been stuck in this forever. Loop of R and D and haven’t actually gotten to a job site. And typically if you get on site, you meet the people and you start to learn about what customers are winning with it, it, that’s like the trifecta. If you check all three of those boxes, you’re likely working with a pretty good robotics partner.

William Bissett: Yeah. Um, I lied. I have one more question for you. Yeah. ’cause I just looked at Harold and I thought about insurance. Um, and so how are, I mean, we go back, we were talking about, uh, it reduces insurance costs, right? Mm-hmm. But it probably brings new insurance costs to the forefront. Yeah. Um, how are you seeing that from a feedback perspective from your existing customers?

Andrew Ashur: Whenever something’s new, it’s naturally gonna be higher ’cause there’s not enough data yet to support, Hey, this is truly safer. So like years ago, getting, uh, an aviation insurance policy for cleaning drones was so much more expensive than it is today. From like a workers’ compensation standpoint, though, in the state of North Carolina, it’s more than twice as expensive to insure an above ground cleaner than it is an on ground cleaner. Uh, so part of the beauty of that is like insurance companies with more data, with more proof of the technology, they realize that, and now our customers are saving as a result. Uh, but that’s not always the initial case. With insurance companies when they’re like, I’ve never written a policy around this before.

William Bissett: Yeah. So, yeah. Super interesting. Jim. My, my mind keeps going, you know, with your cleaning. Yeah. Drones pressure, power wash through my mind keeps going through firefighting. Yeah. Are you in that space at all?

Andrew Ashur: Not yet. And it, it’s interesting, we’ve actually heard like. Two different opinions from the market. Some people were like, yes, we want this yesterday. And other people were like, is it really practical? And to the practical camp, a lot of fires start like, well, within a building, typically in kitchen or electrical, uh, very rarely do they happen on the facade of a building. And from my understanding, I’m not a firefighter. If anybody is here, please correct me.

William Bissett: You weren’t a coder or drug builder either.

Andrew Ashur: That’s fair. That’s fair. But my, my understanding of the way fires are fought is you evacuate everyone first before you start spraying. Otherwise you’re gonna rise the temperatures in the building to a point of killing the people. So it, it wouldn’t do much good in the aspect of like saving lives. Uh, ’cause you’ve gotta evacuate them first. There might be some applications and like fighting forest fires and some outdoor uses. Um, but I, I don’t know today if there’s a great property use case, but for things like, uh, I think hazmat or like oil spills, we’ve gotten some requests for things like that that I think might be much more relevant.

William Bissett: About that for saving lives stuck in a second, third, fourth story or buildings like being able to lift them out.

Andrew Ashur: Yeah, it’s, uh, it’s doable. Technically, one of the big challenges, uh, the drone industry faces is regulations today to operate in, like the standard classification. You can’t fly a drone above 55 pounds, including whatever’s attached to it. So I, I’d like to see that continue to evolve. Because you certainly see drones that are heavy lifting and capable of lifting hundreds and hundreds of pounds. Um, some of the things that would be challenging, just like access points to get those people, uh, but like apartment complexes or buildings that have flat roofs to be able to run up and, yeah.

William Bissett: Yeah. So we’ll go ahead. So with the Sherpa mm-hmm. Data. Yeah. Machine data, like are you, is it ambient temperature, wind speed, and, and all of those other things?

Andrew Ashur: Yeah. Well, we’ll get like a day of data. So certain things like GPS coordinates cross reference it with like, what’s the weather at that time. Uh, now we’re starting to layer in visual data. So just very simple stuff so we can start to train things to understand like what are the abnormalities that are happening on the exterior of this building that might be useful to the property owner. So things like the crack detection or windows being reco.

William Bissett: Um, regulatory speaking, you bring up regulation with the, um, the lifesaving component of it, right? How much does DC and local governments hold back the. Potential development of stuff right now, just based on outdated laws. Yeah. That were there for something else.

Andrew Ashur: Oh gosh. Um, okay, so that answers my question. Yeah. It’s, it’s un it’s unfortunate, uh, in some ways ’cause you, you wish the speed of evolution from a regulatory standpoint matched the speed of the technology. And I’ll just briefly speak to delivery, like delivery drones. The technology’s been there for years. Like it’s ready, but it’s still not commonly deployed because we need more safety cases. We need to evolve the regulations. Uh, and I remember, uh, years ago we looked into getting a, a license to use our drones for agriculture. And after two years we abandoned the process because of how much time and money it was taken. And at that time, there was only one other entity in the state of North Carolina that had it, and it was the N-C-D-O-T. So we would’ve been the first private company to get it. And we were like. Two years in, it just, it wasn’t worth it anymore to us. ’cause it wasn’t the core focus, but it’s like two years for, uh, and, and specifically for agriculture, the legislation we were trying to fit in was FAA, a law that was written in the sixties, like before drones really popular. So here we were trying to like, it, it, it was harder and I don’t know what’s changed now, but it was harder at the time to be allowed to spray pesticides with a drone than it was to actually do it with a manned aircraft.

William Bissett: That’s crazy. Um, and both easier than the getting a real ID. Um, so anyway, so, um, well, Andrew, I mean, uh, again, I’ve been a huge fan of yours for, uh, five years now. In terms of, you know, what you’ve done with Lucid Drones now, Lucid Bots. Um, thanks so much for coming to talk with us today. Super interesting topic. Um, stick around and I’m happy to buy you a beer, um, or a cup of water. Um, I’m sure there’s probably some private questions that’ll come your way as you pack up and get outta here. But thanks so much. Um, thanks everybody for coming out. Um, we’ll do again, our monthly speaker series. We’ll take a break in June and July. We’re gonna do two podcasts with owners that have sold their business. Um, which will be fun just talking about the, uh, the stressful process of selling businesses. We’ll start back up in August with, uh, Culture Index. We’ll have a panel of three Culture Index people talking about personality profiles when making hiring decisions. I, which will be really cool. Um, in September, we’ve got a client actually coming in from out of town, um, who runs a, is it September? Yeah. Um, has a roofing business that, um, has a, uh, a giveaway program that they’ve instituted over the course of the last couple years that. The intent was to help the community and the, um, intent is still there to help the community, but it’s created an enormous referral business as a result. Susan’s gonna come in and talk a little bit about that. Uh, we’ve got marketing, um, tax implications of selling a business and a couple other things as we round out the year, um, all of them. Hopefully Chris will continue to allow us to come back here and occupy his wonderful space. So anyways, hopefully we see you again. Um, stay tuned. And Andrew, thanks so much for coming. I really appreciate it.

Andrew Ashur: Thanks for having me.