Zach Rattner of Yembo: Helping Home Service Companies Use AI

Startups get the party. Startups get the party started. Startups get the party. Startups get the party started. You want tech and tacos, tacos and tech, you know we got tacos and tech is on smash border green guacamole keeper captains counting beans Contador quentan frijoles found us trying to raise some queso la vida San Diego school of rappy rapping fire tortillas round jalapenos in these modern times a crappy and coffee my coding yoga posse fold the body origami we be surfing coronado eating tacos El gordo, the taco standing puesto lucha libre and sombrero Got a shout out to cost angels by the ocean watch the sunset la Jolla shores Entrepreneurs trying to get funded When San Diego Startup Week is popping in the summer Oceanside, the Tijuana esti be full of hustlers from couch surfing to penthouses right on the beach Tech leaders believe stars are within reach we interview and founders who be shooting for the moon SD Tecos with your host, Neil Bloom. Startups get the party. Startups get the party started. Startups get the party. Startups get the party started.

Welcome to the Tacos and Tech Podcast. I'm your host, Neil Bloom, and I'm thrilled to have you here. On the Tacos and Tech Podcast, we interview builders of technologies, of companies, of communities, and we do it weekly, so thanks for tuning in. Every Tuesday, a new interview, and on Fridays we talk about what's new in the San Diego tech ecosystem from a news perspective. So make sure you join us there as well too. We also have a tech newsletter that comes out every other week@freshbroodtech.com you can subscribe to listen to it without further ado. Enjoy this episode. Hey, and welcome back to the Tacos and Tech Podcast. I am your host, Neil Bloom, and I'm thrilled to have on a longtime fellow startup community member here, co founder and CTO of Yumbo, Zach Ratner. Hey, Zach, thanks for being on the pod.

Hi, Neil. Thanks for having me.

Yeah, it's a real pleasure. I think we got to know each other way back in the day of the original founding of Yambo. I think you were still at Qualcomm making the transition of understanding what it would take, the leap to be an entrepreneur. So I'm really excited to talk through that story a bit, but do you want to give us a quick high level of what Yembo is?

Sure, yeah. So Yambo is a software company and we bring digital transformation to home services companies. And we do that using best in class artificial intelligence technology. So our core product or flagship that we started with is in the moving industry. And I don't know Neil, if you've ever used a professional moving company, but the process can be quite tedious, quite complicated. Oh yes, yeah. Traditionally movers have to spend all this time and energy to identify what you have and that's critical for coming up with an accurate quote. So what Yambo provides is software where they can just send a link to the client. The client can click the link, take quick videos of each room in their house and our AI will identify what's there, how much it weighs, and a volume estimate so that you can get an accurate job estimate based on the actual items that are seen. So it replaces these long back and forth phone calls or an in person visit. And more recently we've launched similar products in the property insurance industry. So similar workflow, similar benefits, just new like core AI technology since we're detecting different things.

Really interesting. I have a personal story about completely losing a whole bunch of items in a move. And while I, you know, it was unfortunate, it was very freeing as well in the end. But hate to lose personal items, that's for sure.

Whatever the goal.

Yeah, the moving tech, you know, space has gotten a little bit of attention I feel like over the last 10 years, you know, from like clutter and even some local like tech adjacent service companies like hire a helper as well too. So it's been cool to see the interest in spaces that are not tech savvy starting to get some tech in it.

Yeah, I mean it's a point in time in people's lives where they are very frazzled. There's a lot going on. I don't know many people who go through a move and say everything was perfectly fine, nothing bad happened and historically there haven't been a lot of good technology solutions out there. So we like to joke that. It seems like in a lot of cases Amazon cares more about tracking my like $2 toothbrush order than a lot of these like pen and paper home service companies have. And it's not for lack of trying, it's just there aren't a lot of good options out there in a lot of cases. So that's where we saw an opportunity to do something and make the industry better for both sides of the marketplace, for the service providers as well as the end user.

That's great. And I want to hear that founding story a bit more as well, especially how you met your co founder, Sid. But I really want to get to know a bit more of your background, because I'm guessing you didn't get into tech and being a software developer to, you know, to revolutionize the moving space, at least at first. Where did it all start for you? How did you get into the tech world?

Sure, yeah, it's. I've kind of always been a nerd. I've been writing code since I was a kid. I mean, I started off making, like, silly Flash games for my friends, and I'd burn it onto a cd, and when my friend's birthday party, most people were giving them, like, toys and gift cards and stuff. I give them, like, a homemade video game, and they think it was cool. I had two siblings growing up. Still have two siblings. I just don't live with them anymore. But we only had one family computer, so time was always at a premium. So I remember I would save up my allowance and go on down to the dollar store and get a yellow legal pad, and I would actually write code on the legal pad. So when I'm waiting for my turn to get onto the computer, I could have everything I wanted to put in there in my game. He actually got pretty good at debugging. And a lot of classes I ended up taking in college ended up being kind of similar to that, where a lot of my classmates were complaining, like, who writes code on paper? And I realized that it's kind of funny. Growing up, I actually kind of did. But I just always found the problems that engineers solve pretty fascinating. And it also seemed like from the few engineers I got to meet, like, as a high schooler growing up, that they seemed to have a lot of fun at work. It didn't seem like it was like drudgery. They would always, like, have cool projects where they're working on or working on code for fun in the evenings and the weekends. And it seemed interesting. And then I think maybe two, three years later in high school, I realized it was actually like a discipline. You can go to an engineering school and you can major in computer engineering. So I decided to do early admissions into Virginia Tech. I applied early. I figured if I don't get in, then I'll go come up with a plan B. And I ended up getting accepted. Never had to figure out the plan B. So ended up going to Virginia Tech, computer engineering major, and then studied that for four years. And when I was graduating, it was right when LTE was coming out in the cellular world. So I had an opportunity to move from east coast, from Virginia, out to San Diego, and that was kind of how I landed here in San Diego initially.

Oh, wow. So you moved specifically for Qualcomm, or you were looking to move in general and just come to the West Coast?

Yeah, I mean, I was college kid, senior year, looking for a job, and I had a few different opportunities, and Qualcomm was the best. I wasn't looking to move to any one particular place, but I also did intern at Qualcomm over a summer. And. And as a person who grew up and spent actually five years of my childhood in Vermont, coming to San Diego was a welcome change. So I think I had a really good experience there. And then when the offer came in, it was like coming back to the same team. It was seemed like right place at the right time, like, fantastic opportunity. These generations in mobile wireless technology seem to come around once every 10 years or so. And I was kind of like perfectly aligned right when it was starting. So a lot of things just lined up correctly for me to make that transition.

That's great. Well, obviously one like Qualcomm has had some iterations over time too, from going to 5G and beyond. Now, what were you working on? And I'm curious, just kind of like how that started getting you to think about other things you wanted to build.

Sure. So I joined Qualcomm on a team that works on the signal bars. So if you look at the top menu bar on your phone, you see, I think on iPhones, they're little circles. On Android phones, they look like bars. But when you're going down the highway at say, 60, 70 miles an hour, out of range of one tower into range of the next one. Our team managed all the software on how do you not drop the call? So we'd answer questions like, when is a really strong 3G signal better than a really weak 4G signal? And all the algorithms behind that. But at the end of the day, it was to just maintain the strongest connection to the cell tower. So, like, very fundamental technology. And I was really fortunate to start my career at a place that was shipping at such scale. I remember we'd write code and ship it off to the test team, and you get, like, reports back and dozens of different towns that have been tested in overnight. So if there's a bug in your code, this testing team is very good at finding it. Some of the brightest embedded systems engineers I've worked with got to work side by side with and learn how these processes and things work at scale. Very different than the scrappy startup world, where you maybe have only a Few customers and you're sort of going door to door in the early days trying to get people to use it. This is like the polar opposite of that, where millions and millions of people are using it.

Yeah, that's true. I mean while a B2B product, like a very consumer kind of feedback loop. Right. Like is my call working or not? That's general public there. But you were starting to already get involved in the startup scene. I think I actually remember you being involved on Startup Week. Didn't you lead the, the developer track at some point as well?

I did. Good memory. And yeah, so I was interested in startups always. And when I first moved to San Diego, this might date me, but I gosh, it was 2010, there was not a whole lot going on and I was just impressed in that like next five years, 2010 to 2015, how many things just happened in the area of which you. Austin, New Decker, a lot of like common names that come up in the San Diego startup scene now. Kind of laid a lot of groundwork. But yeah, I was just always interested in like how can I volunteer, how can I hang out with startups, how can I see what's going on in the startup space? And yeah, learned about San Diego Startup Week. I actually didn't volunteer to sign up to be a developer track captain, but the person who did invited me to join and help her out because she was really good at like community organizing, didn't have an engineering background, so I'm glad she did. It kind of worked out well in the end. But yeah, it was just great. I mean we met a lot of, met a lot of awesome folks. A lot of people are in my kind of like circle of like new technology trends come out. What do you think of ChatGPT? Like these people I still message a couple times a week now. A lot of them I met in those early days doing San Diego Startup Week events.

That's awesome. And we were so appreciative to have that. And I think that was our first year we went skills based tracks. I think if I remember right, either the first or second year. So having a dedicated track to developer was a pretty big change I believe for us.

Yeah. And I just remember being super paranoid. I'd never organized an event before, didn't really know the group that well. Everyone seemed friendly. But like I, I didn't have like super deep roots at the time and I just remember like the day it was going to start, I woke up like 3 o' clock in the morning just like, you know, when you wake up sometimes, you're like a panic and you don't know what's wrong, if something feels wrong. And I just couldn't shake the sense that like, no one was going to show up. And it didn't really go away until we had like, I think it was our second or our third event and we ran out of room. There was not enough space. People were like standing in the back. And then I like breathe my sigh of relief, like, okay, these people know what they're doing. This is legit. We're onto something here. But I just remember, like having zero clue what to expect going in. Like, we put in all this effort and a million people could show up, could be nothing or somewhere in between. So we'll see.

Yeah, we get that feeling every year, that's for sure. It keeps us on our toes, keeps us on our game. Right. It's kind of the feeling about your startup as well, you know, people going to like, my baby, are you going to say bad things about it? You know, but you were getting. What was interesting was to what If I remember back, you were working full time, you were getting involved in the startup community and working on a startup idea. So all of a sudden you're very busy, but learning new things, taking risks with your time, because it could have been very easy for you to just go to work and come home and not do that. Did you, I don't know if you remember, but did you feel like this was just part of an education of getting involved in the startup community to help you get educated to start yours or unrelated.

Yeah, I think it was. I mean, it was a bit gradual. I mean, when you're going into college or grad school, you kind of know what you're getting into. I started out wanting to get more involved in the startup community. I didn't know exactly where it was going to end up. I didn't have like a five year plan or anything like that. And what I saw was this was like 20, 2016 was when I left Qualcomm. I did some consulting for a while to kind of pay the bills. And then Yumbo really kind of started becoming a day job, 2017 or so. So there was some interim time where I needed to kind of figure out what I was going to do, but I knew I wanted to kind of become an entrepreneur and I wasn't exactly sure how that was going to pan out. Kind of debated like, do I want to just work for myself, do some software consulting? And tried that for a little bit and it was like, okay, paid the bills, but didn't cure the entrepreneurial itch, if you will, of like building a product and changing an industry and things like that. But I feel like the main thing that like made it all work in the beginning wasn't like a degree or like a certificate or anything. It was just constantly chipping away at small little bits of a problem until you realize that you're onto something. So I could maybe go into a little bit of background on how Yimbo started and like how we ended up going from signal bars to moving industry AI.

Yeah, I love that for sure. The idea, the team member, the co founders, like take us through what we were going through then.

Sure. So I worked at Qualcomm for five years, but after about three, I switched from the signal bar team to an innovation team. And that's when I met my now co founder, Sid. We were looking at some up and coming technologies in deep learning. So I got to go to like a few artificial intelligence conferences, like learn what's going on in the industry there. And my wife was working at a moving company at the time and like I mentioned before, we're both east coasters, so she actually tried to resign because we're getting married and was going to move to the west coast. But they liked her so much they said, hey, we don't normally do this, but why don't you just work from home? She was working in the international division, so a lot of the phone calls were like weird hours anyway, things like that. So it wasn't that big of a deal to be three hours back. So I ended up in this situation where I'm spending my morning working hours like studying what's going on in deep learning, the advances in artificial intelligence. And I come home from work and our living room would become like the back office of a moving company. And I got this direct insight into just what is a day in the life in the back office. What kinds of problems come up? What are the downstream implications if, if a surveyor goes out to a home and maybe forgets that there's a basement, forgets that there's a garage and you send a truck the wrong size. Like all the operational issues that stem from not being able to give an accurate estimate. And this is at the same time where there was this popular academic benchmark called ImageNet. And in 2015, the baseline algorithm that was submitted to these universities, corporations would submit algorithms and you can think of it like a thousand way multiple choice test. So you hold up a picture and say like, what is this of and it could be like a horse. There's a bunch of different species of dogs in the data set. And your goal would be to write an algorithm that would accurately detect what it is and they baseline like humans who went and identified it so you could go see like what percent accurate are you. And in this time, the number one leading algorithm actually beat out a human. So it was like for the first time in the history of everything, I guess computers were better than people at identifying objects and images. And it seemed like relatively few researchers knew about this. And at the same time I'm also seeing all the downstream implications of people missing things when they're looking at things. And that was kind of the initial thought. Didn't really do anything with it. I mean, as you probably know, it's not very easy to train AI and it's like easy to talk about, okay, we'll have AI go do X, but to actually make it real and make it happen, we can talk about that in a bit. There's a lot of work that goes in, a lot of experimentation. And the more we learned about the space, we ended up buying our own home a bit later. And like going through the moving process ourselves, we called a couple local moving companies to go through that. And that's what really like made me convinced that there was a real problem to be solved. Here was this big moving company, international division is having the same problems. All these local companies were calling in San Diego trying to do like a three, four mile move down the road. And that's when we realized it's even not just moving, it's all these home service companies is you need to have a trained expert. Like if you just take some random person and say how many boxes are in the bookshelves in your living room, like you'll get random numbers. You need to be a trained expert. But these deep learning algorithms, at the same time, we're saying if the information is there and a trained expert can look at the video and identify what's there, then an algorithm can do it too. So it felt like the building blocks are all there. And it's easy to have these ideas like, come and go. But this one, like I couldn't shake it, couldn't convince myself it wasn't worth doing and kind of led us down this journey that turned Yumbo into the company it is today.

Oh yeah, I love the personal story meets. I see a tech opportunity here. So how long did it go from idea side hustle to what was the moment you felt we got something here? We Got to go all in.

Sure, yeah. So there were two, two main challenges we were trying to tackle. The first one was we knew it was going to be very challenging to build AI that can do this right. It takes trained surveyors to do this. It's like it's a profession that requires training, it has skill required. You need to kind of know what to look for. And on top of that, you need like a compelling reason for the product to exist to get you until that state. So now our AI can detect over 200 different items, but in the early days, our first one, I think we could do five. So that means if you were to scan like a typical living room, you would see a tv, maybe a sofa, love seat, but no lamps, no rugs, no wall arts. Like, if you're banking on making technology that's viable for this industry, and I think a lot of artificial intelligence companies run into this, is what do you do when something's halfway there? I think Teslas are a good example. They still have this problem of full self driving is not really full self driving. And if it's kind of self driving, but you still have to sit there and hold your hands on the wheel, you can argue about how valuable that is. And I think the key thing that we were able to find was we were still saving you the drive time, we were still documenting videos. So even if the AI did absolutely nothing, there was still some benefit to an end client to use our product. And I think that was key that we landed on that because otherwise you would just be like waiting for this nirvana state that may never come or you may not exist or have enough cash in the bank, like to get to that point. So what we initially did was we built an interface where the movers could review the results that come back. And at first, if we can only detect five items, it would maybe detect a few things in the living room and nothing in the kitchen, nothing in the bedroom. But there was again still a, a reason that someone would be compelled to go through this case is you're not doing, not burning gas, not wear and tear in the vehicle, not having to schedule an in person visit. There's enough benefit there. And that gave us a conduit that we could look at. Where are people spending their most time? So then we noticed it's usually like the more common items. So that's got us the information that we needed to go from five items to 10 to 15 to 20. And every time we did these releases, the base of clients that we're working with would See, the time it took them to do their job gradually shrink. So today the AI can detect over 200 items and we're seeing like a typical sales team can do around like 3 to 5x more surveys in a day than they could without having AI. But that's not something that just like happened overnight. We were probably like 5% better in the beginning, and then we gradually worked our way up to where it was now. And I think it's almost like a common misconception is if you're reading TechCrunch, you just see the breakthroughs come out, but a lot of times you don't realize that there can be years that it takes to go in to actually these things practically work.

Yeah. Wow. A lot of training for sure. And I think obviously ChatGPT and all that's like really popular now. But, you know, AI or whatever the back end is, it's been a series of, you know, algorithms, analytics developed over 30, 40 years really to get us to this stage. Right. And image classification is just like one little piece of that. While training your AI and adding features that would keep people on, like, how did you, what was you, what did you first go to market with? I guess. And what did the industry start telling you we want more of? Was it the image classification in general?

Yeah. I mean, everyone was compelled. I think AI was a shiny buzzword, but it still kind of is a shiny buzzword. But I think the real value of the AI is the promise that brings in consistency, efficiency. People have good days and bad days. AI doesn't always consistent, and the ability to make it a much more seamless process. I mean, if you're calling somebody, asking them for two hours out of their day to walk around the house, like, you may not even get time on the calendar for a couple of days or weeks out. So by shrinking that process down that I'm asking you, I'm going to text you a link, a few minutes of your own time. It's much easier to get that in there. And I think that's one thing that we did do right in Hindi, there are a lot of things that I learned the wrong way by getting to course correct, but I think by not hanging our hat on any one particular feature and by setting up a framework where we were shipping and iterating quickly, we still have a release window every Wednesday at our company, and that's been that way since the early days. So every single week there's something that gets shipped. Now, the bigger features, obviously you can't design, scope, implement everything in a week. But every Wednesday, something goes out. And what we would do is we put together, like a feedback process where we'd look at both qualitative data. Like, the benefit when you're small and you only have a few customers is you can take a Monday, take a Tuesday, block it off, and just talk to clients all day. How's it going? What's working, what's not working? You obviously have to find the right kind of client that knows what they're partnering with an early startup to be able to put that kind of time in it. But I mean, in the B2B world, at that stage, you really only need a couple. Like, you don't have to go find a cohort of 50 clients wanting to do that. Like, you start with two, start with three, and then work your way up from there. And then we also looked at, like, the hard data, like in analytics of how long were people spending. And that kind of guided where we wanted to invest our engineering time in what to work on next. So in the early days, before we had any product, we would just go down Yelp, go door to door, cold call and try to get clients to sign letters of intent, basically saying, like, here's a demo. Doesn't work yet, but at some point soon, hopefully it is going to work. And here's the pricing we think makes sense. And like, if such a thing existed, I'd be interested in buying at that price point. It's not like legally binding or anything. It's just like forcing you to, like, articulate the value prop, seeing people are interested. And then as we got ready, we would, like, take in cohorts of people. And that was how we got the first clients on board. It was actually funny. We had a. There was an industry trade show and it cost, I think it was $900 to get a booth. And I remember this was. I was still just paying out of my pocket at this point. We had no revenue or anything, no employees, no investors. It was like early days. And I remember I did the math and I thought, if I buy a booth, I don't know if anyone's going to show up because I don't have, like a brand name or anything. But I've got this demo video I made on my laptop and if I just cold email all the attendees and offer to buy them like a drink at the bar, I could probably talk to a hundred people before I spend $900. And that's what we ended up doing. So our first. Our first, like, cohort of folks, and we Wanted to talk to not just small, local moving companies, but also, like, interstate ones and larger ones as we went to this conference. And it only cost, like, a price of gas to drive out to the venue and a hotel room for the night back. And I feel like the big lesson we learned there was, like, try to find the most compelling way to unblock the next milestone you're looking for. And I would honestly recommend, don't worry so much around whether that scales or that looks right or it feels right. Like, if you looked around at what everyone else at that venue was doing, everybody had a booth. Like, it was weird to not have a booth. And I think I. I heard through the grapevine through one of my friends, that I think they actually changed the rules now. And you do need to buy a lanyard. You can't just show up anymore. So I'm sorry if that's something that people consider, but I think the point is I was mostly interested in, like, are people interested in this? Or am I just convincing myself it's a cool idea? And if they are, like, what were the main things that they would want to see? So before we wrote any significant lines of code, we were already having a good understanding around. Everyone was saying feedback kind of fell into two categories. There was the group of people that said, this is the greatest thing since slight spread. If you can pull this off, you can totally change the industry. But then for every person that said that, there was another person who said, you're totally full of hot air. There's no way this is ever going to be done. And it was like a very polarizing idea. I thought that was kind of interesting where I feel like if you have a bad idea, you're maybe met with apathy when you talk to people. But if everyone either loves it or hates it, it's like a. It's not a done conclusion or anything, but it's like maybe a sign that you're onto something. And. And I feel like that was just like having a bunch of conversations early on over something that we hadn't sunk a ton of energy into building yet. Just get the conversation going. It's just the kind of thing you could see yourself using. What are the pitfalls you would want to see? There was even doubt then that people weren't always sure that it would be possible to, like, send a customer a link, have them record the video, have them be willing to do it, have the information that comes back be sufficient to actually give an accurate quote. Like, none of those things actually take AI. We actually Proved it just by giving it away to free for free to a few clients and saying, just watch the videos and tell me is it useful or not And I'll pay the cloud bill for the next like month or so. We'll see how it goes. And then if you say, hey, this is super frustrating, there's no way I can actually do my job, then that's like a clue that we haven't quite nailed it. And if people are saying, hey, this is great, can you make it better, faster? Here's a list of features. Like it's a sign that you're kind of onto something.

Ah, I love that scrappiness. I remember running into you. You had an office in Toronto Valley, I think I like randomly opened the door and found you there while I was visiting someone else and saw you guys showed me this giant screen that you had at the time. What was the giant screen for? Do you remember?

I actually still have that in my garage if you ever need it. But yeah, so this is two trade shows later. So we go every year. We bought a booth this time and we had a working product and the biggest objections we were getting is we found when we would. This is still before we had a sales team or anything. So my co founder and I cold calling following our own Drip email campaigns. The biggest objection, we would see where people were, wouldn't be interested in talking with us because of disbelief. They were like, this is like too out there, it's too weird. I don't want to like waste my time and just talk about. I can't like wrap my head around. It's not like anything I've heard of before. So the big push we wanted to make is we have this product, it's viable, it works and it's real. So what we did was we. I was actually at the Westfield mall where the inspiration struck. I don't know if you remember there, but they have the like the big TVs where they have the mall directories on. There's like ads that play on it. And I went on the e commerce site Alibaba and I found a factory in China that would make me a Yimbo branded one. So we had this huge display and what we did was we sent everybody invite codes and said, hey, here's a link in the Drip campaign leading up to the event. Here's a link, click on it. Scan your house before you leave, your hotel room, when you get there, the conference room, whatever, and then come by our booth and punch in the code and we will print out the Actual live inventory list. If you don't think this is real, if you think there's like a team of people in the back office, like identifying all the objects for you, like, I'm letting you call my bluff because you can go scan something and then come back 10 seconds later to the booth. And we made it like a photo booth. It turned on your face. AI companies, we did face detection. We detect where your face was, stick it on the visual inventory. And it gave you like collateral that you could go back to your boss afterwards. When they say, how was the show? What did you see? You can say, I met Yumbo. Here we go. And there was like a takeaway that came from it. Probably one of the most complicated trade show booths I've ever done. There is a lot of moving parts. We had to buy backup blue ink because our printer ran out of blue ink beforehand. So there's a scramble halfway through the day when it ran out of toner. All these fun things were going on. But I think the point was like we wanted wanted something people remember, right? Like you're walking down aisle after aisle of booth and all these different companies. And we figured, do you have like a huge giant iPad? We had people come up to us and say like, are you the giant iPad company? Like it's something that people remembered it by. And now I let my kids watch YouTube videos on was memorable.

I remember it.

That's been what, three, three, four years now? You remembered it pre Covid easily.

Yeah. Where are you today?

So what?

So take us through, you know, where is Yenbo now?

So now Yambo is all over. So we have a team of about 70 globally, probably about half that's in the US and half's in other countries. We technically have headquarters in San Diego. Probably 3rd ish of the company is still based in San Diego. But when Covid hit like a lot of companies, we realized we had to become remote. Then we started recruiting, realized we kind of figured out the remote work thing. We were able to have people working efficiently. We already had an India team pre Covid when we were in the office. So we were kind of used to putting your work in tickets and doing things asynchronously. So that transition, we had like some shoulders to, to stand on to get it started. So we ended up just not going back to having a physical office. So we do like occasional in person meetings. We have co working membership. So people want to like drop by on a day by day basis or whatever they can. But we're yeah, more or less remote. Company now.

Okay. And where's the technology and and like usage is? It stayed within, you said obviously within moving. You've started to touch a little bit of insurance. Take us through like current day product usage.

Sure, yeah. So now our moving product is deployed, deployed over the globe. We didn't really intend on going global, but moving is very heavily networked. Like if you move from New York to London, usually it's one company that handles the US side and one that handles the UK side. So we have data centers in the eu, we have data center in Singapore as well as in the us We've got clients all over the world. Like I mentioned, the product is taking these 20, 30 second video recordings. Every day we do a few hundred hours of video processing. So it's like running at scale. And we just more recently launched some newer products doing similar workflows but in the property insurance space. So that's kind of the new, new thing I've been focusing on SCTO with my time lately is like taking everything that we've learned from identifying items that people have owned and then leveraging the technology that works well there to do things like underwriting inspections. Like if you're buying a new homeowner's policy, how do you make sure you have the right level of coverage? How do you make sure that everything's accurately assessed? All these kinds of things are very similar workflow wise. Like a trained person generally needs to look or there's a estimate that's made based on heuristics. And our view is that like real data based on what you actually see is going to give a better answer to everybody involved. And we want to make sure that things are accurate for both parties. Like on the moving side, when we started we thought that we would help when there's a price dispute, settle the argument because we had like date timestamp photos. Here's what it was. And I feel like in hindsight that was kind of selling the product short. Where we ended up realizing is that we're actually preventing. The argument from happening is that people aren't always adversarial against each other. But if my price was initially given to be maybe $2,000 and then the estimate wasn't that accurate, then after everything I own is loaded up on the truck, I'm being told it's five now, I feel like I'm owed something and I'll be more likely to nitpick and say there's damage here, there's damage there. But if we just give you the right tool that you just say the right price the first time around. You make it clear. If you forget to record your garage, we'll say, hey, are you sure you don't have a garage? It's unusual for this area. We can just give people more transparency. And that argument, it's not like we're providing a value on who won, we're just preventing that argument from needing to happen in the first place. Which I'd argue is a bit better than making one party win.

Interesting. Oh really Cool. And I gotta imagine you're gonna find new industries and verticals to go after this as well. I'm sure you can't go into details and that's okay. What about just like seeing the evolution of tools for AI since you've been building this. Gotta imagine even that. Like yes, classify image, train the model. But have you seen a number of more tools that just kind of accelerating certain development in certain areas up till today?

Yeah, there's a lot. It's almost like I joke sometimes I need like a assistant just to keep on top of everything. Like just I feel like you're just reading hacker news and trying out everything that's listed. There is probably like a full time job now. But I think there's a lot of really interesting things out there. When we started, there wasn't really anything. We had to build our own internal tool for the, for the initial training just because we couldn't find anything that like worked for what we needed. But now, I mean there's a lot of tools out there in terms of helping you annotate data for training. Because I mean these deep learning networks learn based on examples. So it's, you have to give it a lot of examples and show, okay, this is a kind of item that I'm looking to detect. So if you want to make like a cat versus dog detector, someone's got to go and draw a bunch of boxes around cats, dogs and label them. There's even tools out there making the training data can be tough. One thing we learned pretty early on is that people's homes, when they're moving don't look like beautiful IKEA catalogs and things like that. Like you need to actually look at realistic data. But there's even companies out there that make synthetic Data. They use 3D rendering technology. And that's gotten good enough. Where for certain applications you don't even need like real photos, you can just describe the photos that you want. And there's generative AI tools that are nascent. But the promise is there that you could Even Maybe describe the 3D model and get the scene back without having to go and collect hundreds or thousands of samples of like real photos. Just the pace of everything accelerating has gotten kind of nuts in the last couple of years.

Really interesting. Yeah. And now with it maybe going a little more mainstream one, there's a bit more acceptance of it and then maybe more tools being built for it as well. So you know, just expect imagine more and more. So I hear you needing an AI assistant to just keep everything for sure in touch around that. So what's awesome congrats on everything you guys have built. Because who knew that this is where it would get to when you were, you know, starting startup leadership program and everything else. But you've been in San Diego a while now. You know, we've got a pretty legit taco scene. Where do you go to get your tacos?

Fantastic question. So my favorite taco spot is Stars and Stripes on Miramar, kind of at the end of La Jolla Village Drive. So our office was in Sorrento Valley as you mentioned before, COVID and that was my favorite place to go on Uber Eats. They had good tacos, they had good burritos and it was like 10, 15 minute delivery time. So they quickly became like my favorite spot. And it wasn't an event till about maybe a year or so later that I'm driving for something else and I realized it's actually a gas station. But I think it's like fine. I've even like in interviews sometimes, there was a couple candidates I actually took there for lunch because it was so awesome. So at first I was like, I need to pick a different place. Then I realized, you know, that's like what other dining establishments can fill up your car in addition to yourself? Like I don't know too many. So I wear it with pride now.

How utilitarian of you. I love it. You know, one of my favorites for California burritos is sombreros and they have one in a gas station off Convoy. So I hear that like, you know, there is a, there is no specific mold of what our, our taco places look like. It's just about having good food.

So I love that.

I don't think those rats are mentioned yet on this podcast, which is great. So we've got a new addition to the tacos and tech database of tacos, so we'll be adding that. Zach, it's been awesome to hear your guys story. It's obviously the story is still being written, so maybe we'll have Sidon in a year or two to hear the continuation of it. But congrats and thanks for joining us on the Pod.

Yeah, thank you so much, Neil. Have a good rest of your day.

You too. Cheers. Thanks for tuning in to the Tacos and Tech podcast. You can find all of our episodes on freshbridtech.com as well as on Twitter and Tacos and Tech Pod. Also follow Fresh Brew Tech on all of our social platforms and make sure to share and give a review on Apple Podcasts, Spotify or Google Play. Thanks again.

Startups get the party startups get the party started Startups get the party Startups get the party started you want tech in tacos, tacos in tech you know we got it hey Startups get the party startups get the party started Startups get the party startups get the party started Startups get the party startups get the party started you want tech and tacos, tacos in tech you know we got it hey.