SPEAKER_00: So we need some vanguard here. We need some early adopters who want to get early access. So great. Let's get some of those going.

SPEAKER_01: And then the question I would put, this is something I'm loving brainstorming around right

SPEAKER_02: now, are different ways to hack sales cycles in really regulated industries that are slow moving, right? So in that B2B space, I love talking about this and I'd love, I don't think you and I have talked about it. I would love to hear your thoughts on that.

SPEAKER_09: You know, that's a tough one. You know, whenever you go into education, let's say, or you go into housing, as you are, healthcare, these are some of the giant castles that are hard to scale. We've got moats around them.

SPEAKER_00: They've got drawbridge. It's got alligators in there. Got arrow slits. It's not fun.

SPEAKER_11: This Week in Startups is brought to you by .tech.

SPEAKER_12: Say it without saying it. Head to get.tech slash twist or your favorite registrar to get a clean, sharp .tech domain today. Coda. Coda empowers your startup by bringing words, tables, and teams together. Strategize, plan, and track goals effectively with all your valuable data in one place. Go to coda.io slash twist to get started for free and get six free months of the team plan. And NetSuite. The business landscape is very chaotic right now. That's why you need NetSuite by Oracle. Download the CFO's guide to AI and machine learning for free at netsuite.com slash twist. Hey, everybody. Welcome back to Twist. This is Alex. And today's episode is all about founders. We are talking to three different startups that are each taking a very interesting tack on the world. We're going to dig into what they're building, why, and how it's going to move the needle. First up, we're talking to Paid AI, or just Paid if you want. They're a Twist 500 company working in the world of AI agents. But unlike your Sierras or other companies that want to build an agent for you and deploy into your company, this is all about helping the companies doing the actual building. So what Paid does is help companies track how much money their AI agents cost by checking their number of prompts and which models they're using to get a good idea of what they're spending, and then also helps them with AI agent billing. This is the kind of picks and shovels work that often doesn't get quite as much press as the open AIs and anthropics of the world. But if we are going to bring AI agents into the enterprise, into the world of small business, well, Paid could have an absolutely huge business. Think of it a bit like, I don't know, Twilio, but for AI agents to make a very loose comparison. An absolute treat of a chat. You're going to love it. Then we're talking to the founders from iTrucker. Now, the world of trucking, you may not know much about. You may not drive an 18-wheeler. If you do, shout out. But most people here, I don't think, are doing that. So why do we care about trucking? Well, what iTrucker learned is that the world of truck dispatch, freight management, and all the relationships between individual truck drivers, truck owners, dispatchers, and loads is kind of run by phones. It's not very efficient. So what they've done is built an AI tool that helps all the communications be done simply, easily, and quickly from devices and services people already own and use. I love this company. I love seeing technology applied to kind of like flesh and blood industries. You're going to love it. Then to wrap up the day, 10XAI. Now, if you think about climate change from a risk perspective, what does that mean? Well, you might be worried about increased wildfire risk, increased flooding risk, extreme weather in general. And what 10XAI wants to do is help people both make their homes more resilient and also provide good information to insurance companies to make sure that they're doing intelligent underwriting. You don't want to underprice risk and then get your face ripped off. And also, if you own a house, well, you probably want to know what your risk is so you can go about mitigating it. 10XAI is super interesting. It's a company that I'm paying a lot of attention to. I think you're going to dig it. So with all that said, let's dive in and we're going to start with paid.ai. Let's go.

SPEAKER_14: AI agents. 2025 is supposed to be the year of agents, or at least the year that kicks off agents as

SPEAKER_17: a common tool for companies to help automate their work. But while companies are working to integrate agents into their workflow to try and make more money themselves, what about the companies building AI agents? Are they making money? We know that major foundation model companies like OpenAI and Anthropic are seeing the revenues soar, but at the cost of a huge burn. So the question really does matter. One startup has its eyes fixed on the question of AI agent economics. That company, paid AI, wants to better track AI agent costs and help with monetization so that the work of making autonomous AI tools can prove profitable for both the provider and the customer. To tell us more about paid and the state of the agentic economy, please welcome to the show. It's Manny Medina, co-founder and CEO of paid and former co-founder and CEO of Outreach. Manny, hey, how you doing?

SPEAKER_19: I'm awesome. Good to see you. Thanks for having me.

SPEAKER_17: You get 10 points for having your logo behind you in neon signage. That is fantastic.

SPEAKER_12: I have to ask, how long after the company was founded did you guys get that made for your podcast studio?

SPEAKER_20: So the company was founded around January this year.

SPEAKER_22: We broke code in, you know, towards the end of January, February. And, you know, we've been rocking since. We have, you know, we started with pilot customers and that we now have close to 20 of them. And we just landed our first enterprise customer last week. So we're super excited about that. And the neon sign was a detail that the team brought together. It turns out they're very easy to make.

SPEAKER_24: I didn't know how easy it were to make, but they're fantastic.

SPEAKER_21: It turns out there's actually just a bunch of services that'll make them for you.

SPEAKER_12: But it does make you appear to be incredibly legit. I want to get into customer growth and the exact thing you're building in a second. But I want to just set some foundations here for folks. People have said, like I said in the intro, that it's the year of agents. What I'm curious from your end, Manny, is how many companies are really in the market today that are building AI agentic tools that they're taking to customers and therefore, you know, need what you guys have built. But how many companies have reached that level of maturity, if that makes sense?

SPEAKER_22: So the agent market is incredibly early. And it's incredibly early in that anyone can build an agent, but very few can build an agent that is solving a problem for somebody that is willing to pay for it. But that said, because agents are actually being sold, the majority of the agents that we're seeing are being sold into non-tech spaces. You're not going to hear about them. You're going to hear about them when, you know, when your plumber tells you that it was the agent that they scheduled their appointment and it was the agent that told them that he needed a part. Or whenever your architect uses, you know, an agent to figure out the cost of the house you're about to build or, you know, that, you know, mortgages are going to be performed completely by agents over the next, you know, in the next year or so. So it's a little bit like taking the red pill that once you take it, you can't unsee what's happening out there. And we all here took the red pill and our agents in every single nook and crown of the economy, they're just getting started. Some of them are not yet quite monetizing. Some of them are just, you know, out there, you know, getting market share. But the great majority of them are fine tuning the agents to work precisely well in the industry in which they're deployed, which is incredibly exciting because it's opening an opportunity for people who are not traditionally in tech to be in tech and make money.

SPEAKER_12: To use technology tools. Yeah, no, that makes good sense to me. Do you think we're going to see then agents make their way into technology companies more

SPEAKER_28: generally this year, next year? When does the other part of the economy come on to the agent side of things?

SPEAKER_30: I think that usually the way the economic cycles work with innovation is that, you know,

SPEAKER_22: software startups buy from other software startups and you go up and down one-on-one and that's the majority of your market and you try to expand from there and that's when you cross the chasm. I'm seeing it all the way around. I'm seeing, you know, real American, you know, real business companies that are already jumping directly into agents. They're bypassing the whole cloud thing and going directly into agentic purchases and companies that you will think have no business being an agent, such as freight forwarding or, or, or supply chain or fleet management or mortgages or insurances. They're all going directly to agencies. And there's two reasons why this is happening. One is because their workforce is aging out and they found agents as a way to, to not have to go and retrain, you know, new kids to become actuary who they may not want to be. The other one is the fact that the, the comparison of an agent is not against other agents or human beings, it's against BPO's, which are human beings, but they,

SPEAKER_33: BPO's are business process, uh, outsourcing jobs, people that will do the busy work for you off-seas or overseas, I mean.

SPEAKER_22: Exactly. Exactly. And, and an agent compared to a BPO is incredibly more efficient, more effective and higher performing. So that's sort of like the table stakes right now of what's going on with agents. And that's a very rich market to go after. And we're going to see that for the next five years until we start seeing other effects of agents in the marketplace.

SPEAKER_39: Founders know you can't plan a future strategy for your business without having the most up-to-date and accurate information possible. Listen, nobody's got a crystal ball, but you can start to future-proof your business by using NetSuite by Oracle. It's the number one AI cloud ERP on the market. What's an ERP, you ask? Enterprise resource planning. It brings every facet of your business into a single unified platform. That means your accounting, your financial management, inventory, and more in one fluid, perfect platform, giving you the control and visibility you need to make rational data-driven decisions. And if you're running a complex global business, NetSuite's one world solution makes it easier than ever to track all of your most important metrics. You can keep an eye on multiple subsidiaries at once and automate your intercompany transactions with NetSuite's comprehensive AI powered approach. This team has over 25 years of experience helping thousands of companies from every industry grow and thrive. Here's your call to action. Oracle's got a free gift for all of our Twist listeners and viewers. Download the CFO's Guide to AI and Machine Learning for free at netsuite.com slash twist.

SPEAKER_40: Once again, that's netsuite.com slash twist.

SPEAKER_12: So agents are here today, often in places where you can't quite see them. You're optimistic about their ability to improve and expand their market share. So this is the year of agents year one, I guess. I wish to think of it, not that we're going to get to the end of this year, but certainly we are accelerating towards an authentic future. Fair enough?

SPEAKER_22: A hundred percent. You just planted your asparagus field and they're all sprouting. So all you're going to see is sprouts.

SPEAKER_24: Next year is going to be a nice, bountiful harvest.

SPEAKER_12: Okay, let's talk about where Paid fits into this because everyone knows in a traditional SaaS context, software is very margin efficient. You can have gross margins in the 70s, 80s, some cases even low 90s. One reason why everyone loves it. AI, on the other hand, is often more expensive to run. And I presume that when you build an agent that has a foundation model from a third party inside the core of it, it's expensive to run as well. So can you just tell me what the economics look like for the average company selling agents for those products? How much do they cost to run?

SPEAKER_28: And if you can, before Paid came in, how are they tracking those costs?

SPEAKER_22: So the difference between a normal SaaS company, like a SaaS company of the 70 to 80% gross margins and an adjuncting company is exactly what you talked about. Is it the cost of inference and the cost of tokens that are paying to the foundation model provider and those go on top of your cloud costs? So now you have your 80% cost and you have the additional cost of, you know, pretty much the human replacement. So whereas a SaaS company will have, you know, 70 to 80% of gross margins, an agent company would have somewhere between 40 and 60% gross margins. And that will significantly, and then that's problem number one or sort of situation number one, situation number two, and that is incredibly variable depending on the type of customer. So for instance, you can have a customer, if you're in customer service, you can have a customer that, that your calls in and you quickly, you know, tell them the answer to what they're looking for and on they move, or you can have a customer that calls in and they tell you the story of their lives. They ask you a bunch of questions that are not relevant to what they're trying to get resolved. And then they ask you the question about, they're trying to get resolved and you just got a free chat session with that person. And that costs you a ton of tokens.

SPEAKER_30: And you're paying for all the tokens. Yeah, exactly. You're paying for all the tokens. You're paying for all the tokens. Exactly.

SPEAKER_12: So really, the more terse versus verbose your customers are, if they're chatting with your agentic product, the more money you can make. So really that does imply that you want to get to value very quickly in those interactions to prevent excess token generation.

SPEAKER_22: A hundred percent, but there's no way for you to predict it. The customer, the service is out there and it's performing. I mean, the same thing is for agents that do document reviews. So there are some people who have short histories and some other have very long histories. And for the very long histories, that's just going to consume a lot of tokens. The same thing for people who are, you know, performing complicated tasks. The more agentic the behavior is, the more autonomous, the more agency an agent has, the more likely it is to run off the rails and go do something that you are not expecting it to do. And it may be good for good reasons, but you just don't know. And that's sort of the trade-off that we're living in right now, which is a fascinating thing because you're giving them the ability to decide. And that's very exciting.

SPEAKER_12: It's very exciting that they can decide, but you don't want them to decide to spend your entire checking account on, you know, anthropic credits or whatever. A hundred percent. Okay. Now, a question that I have is, do some agentic systems use multiple AI model providers? Because when I was thinking about the problem you guys are solving, it seems like it would get even more complicated if I sometimes lean on GPT-5, sometimes on Grok-4, and sometimes on, I don't know, Claude Opus 4.1.

SPEAKER_22: They do, but they don't do it for economic reasons. So if you think about every single model from every single provider, they have different behaviors. They have different tendencies. They are, they may answer the question completely differently, even if they're provided by the same vendor. So you don't know, because it's all a black box, you know, whether you can, you know, hard replace, you know, Sonnet with, you know, any of the versions of GPT and whether that's going to work for you or, you know, Gemini, et cetera. So there are definitely ways for you to arbitrage the token cost, but the tasks need to be relatively simple and relatively swappable. In which case you may not even be solving the right problem because there will be an agent company out there that will take your problem, bundle it into a bigger basket of problems that are solving it and make you irrelevant. So it's a very dynamic environment right now where there is no free lunch for anybody, right? It's like, if you're making a ton of money, you can rest assured that there is four YC startups that are going to come after you with a much better model that is going to solve a bigger problem you're solving in your roadkill.

SPEAKER_33: So it's the old Bezos quote, uh, your margin is my opportunity.

SPEAKER_30: Oh, a hundred percent, a hundred percent at Amazon, you know, people used to live with

SPEAKER_22: like somewhere between, you know, 6% and 20% operating margins. And it was a good day when you hit 20% because people bought a lot of books, but on the day that they bought a lot of electronics, you could be upside down. So you have to be really careful how you operate. This is why, you know, most of the companies give you free coffee at Amazon. So you got to put your 25 cents in the little jar before you put yourself a cup of coffee.

SPEAKER_33: I, I didn't know that, but that was my experience at, uh, the headquarters of state farm. Once I got there after going to a lot of Silicon Valley offices and I went to the little cafe and they're like, that's a dollar. I'm like, what?

SPEAKER_59: There you are where you're, where you're Silicon Valley entitlement. We're hoping for a free latte. And that's a use.

SPEAKER_62: It wasn't even a latte, man.

SPEAKER_12: It was just a black coffee, but that's what you get for going to Southern Illinois. Anyway, that's what you get. Okay. So clearly you don't want to undercharge for your agentic work because then you're upside down and that's terrible. So talk to me about how paid helps people actually sort out the discrete costs per, I'm not sure the phrase here, agentic interaction, agentic use case, the, the, where the rubber meets the road, if you will.

SPEAKER_22: Yeah. So the, if, if you think about the dynamics of pricing, pricing has roughly three components. One is your cost. So you clearly want to be above that. The other one is what is the equivalent of the service that you're providing and how much would that cost? And then in the iteration that we're living right now, that equivalent for the agentic cost is a human being. So you can easily compare, you know, what would a human do take to do the same work? And the third one, of course, is a competition. Like there, if somebody is doing the same thing that you're doing and they have a different cost, you have to be solely in the same ballpark of that. So what we help everyone see is the fact that, you know, A, we help you see your cost, but B, we help you understand what is it the agent is doing so that when you pitch it to your customer, the customer can figure out out of all of these activities the agent is performing, which ones do they attach the most value to? And you can charge for those. And this is the change in mentality that we are embracing. We're actually asking our customers to embrace is that you don't need a consistent price sheet for all your customers. What you do need is to make sure that your value aligns to the customer perceived value receive. Once those things match, that's where you should price. You shouldn't price in a way that is simple to you. You should price in a way that is consumable by your customer in a way that your customer perceives they're getting value. And that's what we offer. It's a flexible platform that understands what the agent is doing and allows you to capture the value in a way that the customer wants the value to be captured.

SPEAKER_14: Okay. So that's, that's two parts.

SPEAKER_12: Let's stick with the, uh, cost tracking. I've, I was going through some of your developer docs and you guys talked about, um, signals as a concept of like breaking down work. I guess what I wanted to know is how granular can you get on tracking, um, either an agent for a customer or an agent doing a task for a customer? Can you get down to the sense that it costs?

SPEAKER_20: So we, what we do is just to get technical here for a minute. So we, we, we found out that the entirety of what you need to understand of what the

SPEAKER_22: agent is performing is in the runtime code of the agent. So once the agent executes his code, there is this egress that comes out of the code that is the runtime and that could be captured by a technology that is open source called open telemetry. Open telemetry allows you to decorate the code and then see the entire egress of the, of

SPEAKER_30: the execution. The egress of the execution is like, imagine somebody is walking around right next to you, clocking every activity that you do and taking a note. Every time you drink coffee, I clock it. Every time you do a podcast, I clock it. You go take the kids, I clock it. And now you have a catalog of everything that you did today. And now you can choose what are you going to charge for and what are you going to show? Right. And that's what other technology does.

SPEAKER_22: And once you do that, then you can go around and buy customer. So if a customer bought your agent because it saves them time, if a customer bought your agents because it, it, it, it shortened times to value to their customers, if you customer bought your agent because it delivers higher NPS, you can charge for those things because

SPEAKER_71: you can see it.

SPEAKER_12: Okay. So you can get very, very granular and you can decide what to charge for. Are you pushing us towards a world in which each customer not only pays a different price for what they're getting in an agent context, but it could also vary based on the specific tasks that are put into it in that moment.

SPEAKER_21: Is that responsive to the customer asks of the service?

SPEAKER_30: Absolutely. So let me tell you an actual exact example. Okay.

SPEAKER_22: Um, we're working with, um, customer service agents for car dealers, and it turns out that this car dealer customer service agent, uh, is blowing up in the middle East and, and the, and the, and the work perform is the same. So they take an inbound lead. They convince it to come into the showroom. They get more information about the kind of car they want. And once they become a customer, they sort of try to get them in for services and for upsells, right? So it turns out that the same customer, the same activity, sorry, not the same customer, the same activity that the agent does. If that person, if that activity is performed in South Dubai and you get a customer in, into the Beamer dealership, the likelihood of they walking out with a car is in the 90 some percent. Why? Because they have a lot of money. If they like the Beamer, they walk off with it versus North Dubai, where the likelihood of they walking out with a Beamer is a lot lower because there's then to be a lot cheaper, a lot more price conscious, and then go from dealer to dealer to dealer to find the right price. So driving a lead into the store at the South Dubai dealer is worth significantly more to the dealer than it is driving a lead to the North Dubai car dealer.

SPEAKER_73: And how do you capture that value? Even though the work is the same, the value is significantly different. You want to capture that.

SPEAKER_75: Founders, you got a lot on your plate. You're running a company and that means you have endless tasks to get done. And you're juggling a ton of priorities all while trying to hit your KPIs. That's why I love Coda for keeping everything under control. It's the all in one platform that consolidates your documents, your spreadsheets, your apps into a single scalable workspace with a beautiful, intuitive UI. Coda helps to find planning sessions so you don't lose focus and it creates flexible agendas so you can adapt to shifting priorities. It's going to help you set your OKRs and you get access to hundreds of templates for tasks like setting priorities and mapping dependencies. We use Coda every day at Twist for important projects like the Twist 500 and Founder University. It's truly an all in one hub for startups. Head over to Coda.io slash twist right now. And you're going to get six months of the team plan for startups for free. That's C-O-D-A dot I-O slash T-W-I-S-T to get six months of the team plan for free.

SPEAKER_76: Coda dot I-O slash twist.

SPEAKER_12: But if I'm the dealership, I want to pay as little for this as possible because I want to make the most money that I can. Am I going to be content with you coming to me and saying, all right, listen, May, if it's a South Dubai IP address, then we're going to charge you X dollars. If it's a North Dubai address, we're going to charge you Z dollars. And I'm going to be like, wait a minute, it's the same work. Why do I have to pay more for it? Am I not just padding your margins? This feels like search pricing in the micro sense for each customer interaction, which feels almost egregious to me in a way.

SPEAKER_28: Maybe I'm just being too cheap here. But Manny, I feel, are people complaining about that at all? Or are they fine with it?

SPEAKER_22: No, I think that it drives a conversation around value again. So if you're a dealer, let's take a step back. If you're a dealer, not an influencer or a podcast or host, like if you're an actual dealer, your job is to move cars off the lot. Yeah. Full stop. Right? And you make money by moving certain cars of cars off the lot. So whether the agent charges you more for, because you happen to be the South Dubai one versus the North of Hawaii one, it may come into play in the conversation. But at the end of the day, if this agent moves cars off the lot, we're having a conversation and there's going to be a productive conversation.

SPEAKER_12: Okay. So I was really thinking too small things. I was thinking about SaaS pricing, you know, recurring per seat, whatever people moved into in the Twilio era, you know, pay for what you use. In this case, though, we're kind of moving to a new iteration, which is pay for demonstrated

SPEAKER_33: value, I suppose. Exactly. Okay.

SPEAKER_22: Exactly. And just think about that. Like if, if you are charging for the, for the fact that you moved the car off the lot and you are not charging for the 20 other conversations that didn't result in moving a car off the lot, there's a lot of economic unlock in that kind of pricing. I'm not suggesting that that's the way you go, but I am suggesting that that is well more aligned to both parties and how both parties can realize value than some kind of seat pricing in which, you know, you may have a person using the tool that may or may not move cars off the lot.

SPEAKER_12: So you just said that when you manage a transaction, when you pull off a transaction via an agent, it's a success, but you might not get paid if it doesn't lead to a sale. So are we going to see a world in which some agentic products only derive revenue when a customer completes a transaction on the other end and then otherwise it'll be provided for, for free?

SPEAKER_27: Yeah, that's it. That's entirely in the case. Wow.

SPEAKER_12: That's super aligned. I think, but also it sounds very complicated given you'd have to have so much information about, you'd have to know that you're, you know, there's a split in Dubai. You've dropped the line here and you'd have to probably define in sort of some billing system. So has all of that technology been built that we'll need, or are we still working on building

SPEAKER_31: the scaffolding up to provide the granular information you would need to be so detailed on your pricing? No.

SPEAKER_30: So that's what we do. That's exactly what we do. So we, we eliminated, so we, we, we built and paid from, from the principles of what are

SPEAKER_22: the future interactions going to be like. And once you decide what the future interactions are going to be like, then you need that kind of flexibility. Yeah. You need to be able to provide it. Like, again, I'm not saying anything new here. Like you should be able to provide the value and deliver the value that your customers cares about and then charge for it. But it's, but as it turns out, every customer is in a different, it's in a different path. It's in a different value perception curve, and they should be able to pay in the way they want. And like what we have done in the last 20 years is convince the software vendors that because we're so powerful and so magical that we can charge this arbitrary per seat tank, because it's really hard for us to like otherwise scale the pricing. And I think that's a big bag of lies. I think that in the end of the world, software is very cheap to build and you should be able to build exactly what your customers need. And like your inability to scale custom pricing doesn't seem like it's your customer's problem. That seems like your new problem.

SPEAKER_35: You should really fix it.

SPEAKER_12: Yeah. Okay. So in the future world, then let's say that I'm selling an agentic product to a Dubai dealership. Do I have like a forward deployed engineer that's helping them, you know, dig deep into their granular business to define the right pricing setup for them? Or is this something that's automated enough that a dealership could set up a relationship with an AI agent vendor without having to do a lot of custom coding work to make their iteration function? Because I feel like if it's the latter case, your business is going to be much bigger. If it's the former case, it's going to be slower growing.

SPEAKER_22: I think that a lot of these rules are going to be baked in. So like as, as one of the things that is beautiful about this, um, agentic world is that, is that we're truly embracing services because we're replacing services. But as you embrace services and replace them, the kind of service you provide is higher level. You see what I mean? So instead of like doing forward deployed for, you know, hooking up the agentic workflow that is going to do discriminatory pricing based on your location, I think you will, you will do some other kind of services around, you know, you know, where car dealers actually make money is on the trim, it's on the services, it's on the upsell. So like the idea that then you, you're going to really have your juices flowing more into like, okay, so what kind of marketing program can I run to get people back in here? You see what I mean? Like, how can I elevate the trim? The moment you fall in love with that red beamer, how can I, how can I sell you the one with the spoiler and, and the one with like, you know, more horsepower. So those are the kinds of things that you're thinking about that, uh,

SPEAKER_27: that this, this kind of like, you know, pricing on locks.

SPEAKER_12: Well, I was really excited to talk to you because I feel like you were attacking a really interesting problem point in the AI economy today, figured out what it costs to run an agent and also how much to charge for it. I didn't realize though, Manny, and this is so glad you came on, just how detailed where you're going to need to get, and also how good that's going to be for the end customers. But what this does do is put a lot of pressure on the AI agent companies, because if we're going to charge only for really defined value, they have to build high quality products that really get there. So I guess, I guess good luck everybody building because there's gonna be no, you have to get signed up for a thousand seats to get the enterprise, you know, you know, SSO or whatever, you're gonna have to really be good.

SPEAKER_95: A hundred percent. A hundred percent. But like, isn't that the way it's supposed to work?

SPEAKER_73: I mean, so why would you go to a doctor that is okay versus the doctor who actually fixes your

SPEAKER_12: problem? Uh, the website is paid.ai. Manny, thank you so much for coming on. We really appreciate it. And, uh, when you hit a hundred million ARR in like 40 months, just, uh, call me up and, uh,

SPEAKER_108: we'll talk again. You're right up. Good to see.

SPEAKER_17: All right, Jason, speaking of founders, we have one literally waiting in the wings. I'm incredibly

SPEAKER_12: excited to bring Camilo Ramirez. He's the co-founder and CEO of iTrucker. Now, you know this company, Jason, it's based in Pittsburgh. It was in launch accelerator 34, and it's working on helping trucking companies do more using AI agents. And what's cool is Camilo is not just another Stanford graduate out there trying to sling code to a market. No, he's actually driven trucks on a cross country basis, both dry and as he says, reefer, which means refrigerated. Camilo, welcome to the

SPEAKER_00: show. Hey guys, thank you. Appreciate it. Yeah. Nice to see you. I was absolutely in awe of your

David Friedberg: performance. You did a great job, uh, presenting in the accelerator and yeah, sometimes domain expertise is great. Sometimes domain expertise can ankle you, but sometimes it can give you a unfair advantage in the space. So what have you learned in that aspect and maybe start with telling us what

SPEAKER_120: your insight was as an insider of what needed to be built here. That'd be a good place to start.

SPEAKER_122: Sure. Yeah. Well, thank you guys, uh, for having me here. Appreciate it. Uh, yep. Former truck driver, uh, a couple of years ago where I understood the pain points of the driver when you're driving over 600 miles per day. Uh, then you're stuck with all, you know, flat tire, uh, at the midnight and no one is there to, to help you. Right. So then we had the opportunity to own our fleet. Uh, we ended up with around like 15 trucks where we saw the problem from the other side. So at the time we had, you know, 15 drivers calling us, uh, asking for pickup number. Then we have 15 customers asking like, where's my load, you know, like, is it going to be on time and it's a 24 seven business. So we understood that, uh, in order to build something to, to, uh, improve this kind of like activities, coordination on a single basis. Um, we needed to build something that is able to be part of the operations team. So basically that is able to talk to the drivers, to talk to the customers, to brokers, uh, forward information. We have a ton of paperwork going around. So, you know, reading those, uh, documents and forwarding the right information to the right party. So we learned, uh, that we needed to build this, uh, bottoms up in order to, uh, get to the,

SPEAKER_124: to the trucking companies, um, put our service within their operations.

SPEAKER_39: Y'all know I'm obsessive about a good domain name and I'm so proud of my exclusive collection of single word, easy to spell.com from begin.com, inside.com, mahalo.com, aday.com. But the truth is, the supply is dwindling. You could spend your whole life brainstorming ideas, but 99.99% of the best.coms are taken by rich guys like me, or they're sitting there wasting away unused, but fear not founders. There's a simple solution dot tech domains dot tech domain means you don't have to compromise or give your company an awkward name. You hate dot tech instantly lets everyone from customers to potential investors know you're a tech company building something innovative. It's in the domain. That's why over 500,000 founders have collectively raised over 5 billion in investment using their dot tech domain. Even CES, the world's largest electronic show is using a dot tech domain. Now for a reason, it's time to move over to dot tech. So head over to get dot tech slash twist, get dot tech slash twist.

David Friedberg: So when this was done prior to your product, iTrucker, which is I-T-R-U-C-K-R, drop the E,

SPEAKER_39: for elite. You found there are dispatchers. Dispatchers are curmudgeonly humans who have to get on the phone with all these parties and yell at the drivers and manage the people who are, you know, sending a load across the country and they're chain smoking Marlboro lights in some boiler room somewhere. And they're getting paid a large amount of money to do this, essentially air traffic control for

SPEAKER_122: trucks. Yeah. Where do they get paid? Yep. So there are like two different ways on how that operates. So we have like some fleets there. They are able to like bring those dispatchers like in-house team of people. So it's a little bit cheaper for them because you're hiring like a full-time employee. But then there's a lot of trucking companies that use like dispatching services. Got it. So those fleets are able to, well, they're not able to, you know, bring a full-time employee. So they're paying like a portion of the revenue of the- Oh, that sounds expensive.

SPEAKER_133: Yeah. It's pretty much expensive. If you have like 10 trucks, you know, you're making-

David Friedberg: So what was your solution then? I know you had an AI-first solution to this. So what did you build

SPEAKER_122: and how has it worked? Yep. So we are deploying AI workers, AI fleets within those operations. So basically, we are automating the communication. So now a trucking company, instead of having like three dispatchers to dispatch a fleet of 15, 20 trucks, they probably just going to need one because we help them with like booking loads, right? So now our agents are able to go to different load boards, find different option of loads, make the phone call, send email,

SPEAKER_138: book the load, then- And that's being done by an AI agent that's calling on the phone using voice or text to voice, I guess. And you've, you've built all these sophisticated AI agents

David Friedberg: and they occur, uh, as a co-pilot where they're supervised or are they stable enough to deploy them and not have a human in the loop or a human checking them? How does that work? Because we've heard a lot with this MIT study, oh my God, you know, AI doesn't work. Obviously, co-pilots work really well. Agents, people do have concerns. So yeah, unpack that piece.

SPEAKER_122: Yep. So we right now have, well, they have a dashboard where they can see like the whole operation. So they can see like how the AI is talking to the drivers through text messages. They can see like how the email is replying on a, on lifetime. Then they also have the option to, to like stop whatever the AI is doing or the AI suggests a few things. So let's say just to, to give you like a specific workflow. So they have a truck idling for six hours. So the AI notified the dispatcher or the fleet manager saying, Hey, we have these drivers in this location, uh, it's idling his truck for five hours. You want me to find a load? So they just approved the, the, the option and then the AI go find the load, book the load and notify the driver and notify the dispatcher saying like, Hey, I found this based on the metrics that we already have set up. Camilo, how do you get that information?

SPEAKER_12: Is it IOT that lets you know that this truck is idle for five hours? Cause that's a very specific actionable data point, but you can't, you know, have a human looking at GPS data. So how do you

SPEAKER_122: get that, uh, that nugget? Yep. So in trucking, there's, um, uh, a system called ELD, which is an electronic logging device where every single truck needs to put a hardware into their trucks to see where they are to, uh, for this hour of service that the driver can drive. They can just drive 11 hours per day. So they have this GPS tracking, uh, kind of thing. So they have all these telematics, like, you know, field consumption and all that. So we are integrating, uh, those ELDs into our system. So we are able to see where the truck is, uh, how fast they drive, if they have, you know,

SPEAKER_146: enough fuel to get to delivery. So truck drivers have no privacy. They can't stop and like have a

SPEAKER_138: couple of beers and drink, uh, or have a hamburger and take a nap and then go 75 miles an hour the next morning to make up for it. They're monitored like very granularly. Yeah. Correct. Yeah, exactly.

SPEAKER_148: So are they required to have cameras in the, in the cabin too? Right. Yeah. Some of them,

SPEAKER_122: they put like cameras in front of the dashboard. Um, is that required now? Or is that like your trucking company decides, is that like a federal mandate? No, no, no, no, no. That depends on the, on the fleet and then depends on the insurance company. Some insurance are like, Hey, you gotta record every single, you know, piece of the, of the journey of the driver. But, uh, some of them, they're just,

SPEAKER_120: uh, you know, fleet. Which is why we now have this video tragically. I don't know. You must have seen

David Friedberg: this of this. Did you see this last week, Alex? They have video of a lunatic truck driver who decides to make a U-turn on the freeway. And I think he had a double cab. I'm not sure if it was a, it was a double cab? Yep. It was the same. It's so tragic. He decides, Oh, I missed my exit. So I'm going to use the

SPEAKER_153: U-turn, which is like the police U-turn. Yeah. In between the dual carriageway, the gravel bit

SPEAKER_154: in between. Yeah. Yeah. And then he, because he decides to do this, a car can't stop and smashes into the side of him. And I think tragically two people were killed. I'm not going to play it

SPEAKER_39: because of that, but no, we don't need to play it, but it's awful. It's awful. And there's a video of his perspective on it. Cause you have not only video of the driver, not only video of the road and

SPEAKER_159: everything. And the review mirror, you have video of the driver and their reaction to it.

SPEAKER_12: Oy yoy yoy. Yeah. That was pretty tough. Camilo, what's your average trucker? How many trucks do the average customer review guys have? Cause what I learned prepping for our chat today is that trucking companies come in size from absolute pipsqueaks to absolute behemoths. Yeah. So we started

SPEAKER_122: with smaller fleets, like one, three, four trucks operations. Once we like fully deployed the product, we now are targeting like midsize fleets. So anywhere from like three trucks to a hundred trucks. Uh, those are the ones that are feeling the pain that they're having, you know, like four or five dispatchers, uh, working 24 seven, talking to drivers, talking to brokers. So those are the,

SPEAKER_165: the, the fleets that we're targeting right now. And we are focusing on how many, how many of those

SPEAKER_12: companies are there in the U S I don't think a lot of folks that are listening to us right now have a good feel for how many midsize trucking companies there are out there. I presume it's a lot.

SPEAKER_122: Yeah. So let's, let's say that in total, we have over like 2 million fleets that includes like box trucks and all kind of like, uh, fleets then within the trucking, we have around like 800,000. Right. And then within those 100,000, we have around, around like 300,000 for midsize, like three to a hundred trucks. So there's like a, uh, plenty of, uh, a room for us to play here.

David Friedberg: Awesome. Well, it's office hours, company's doing well. You got through the accelerator, uh, you got customers, you're growing nicely. Uh, you're using AI to the fullest of its current abilities. You're one of these companies that's like the tip of the spear and using this technology. Uh, but you know, we're here and there must be some things you're either struggling with

SPEAKER_169: or challenged with any questions for me that, or anything you want to discuss?

SPEAKER_122: Um, well, now we are seeing a lot of opportunities like an enterprises accounts, uh, kind of like larger fleets, uh, even like out of the country opportunities. Great. So, um, in order for us, you know, to stay focused on, on how we're growing because we're growing pretty fast, uh, any tips on like how to handle both things, like, you know, like how we handle enterprises.

David Friedberg: Yeah. So the great thing about enterprise customers is that, uh, they will spend more money

SPEAKER_138: and if they love your product or service, they're probably less price sensitive. The great thing about the small boutique, you know, smaller customers is, uh, they might be more scrappy and willing to try

David Friedberg: something, right? Cause they don't have the budgets and you know, that's why a lot of people will use startups as their first customers because startups will actually use the product. Whereas if you go after the big enterprise clients, they take a long time to sell, they're going to need SOC 2, you know, so you're going to need to get in touch with Vanta, make sure your SOC 2 is compliant, all that good stuff. Um, and they just take forever, right? They might have a six month sale cycle. They may want you to build custom software so they can be a bit of a headache. What's important is to study your customer base. And if you're in a highly fragmented market, uh, there can be many customers. What's a highly fragmented market, uh, startup that's done incredibly well? Spotify. They went

SPEAKER_39: after small merchants. There was Magento and other big products, you know, going after the large enterprises going after Target or, you know, I don't know, Walmart, maybe those companies are even too big. They build their own it because it's so critical for the user experience, but they went after

David Friedberg: the long tail HubSpot. They wanted, you know, small, uh, firms, Squarespace. There are big companies that use it. There are small companies that use it. So then you have to make a decision. Do you want to

SPEAKER_120: have a sales team? Do you want to deal with custom software? Do you want to deal with approval

David Friedberg: processes? Uh, and both can work. Obviously there are Salesforce and other tools for large enterprises for sales and marketing functions. And then there's HubSpot. Um, and this will inform a lot of product decisions and a lot of go-to market decisions. Your go-to right now, I'm assuming is founder led sales. You email a small trucking company, or they hear about it from another small trucking company. You get on the phone with them and the person you're talking to owns the business, gives you their credit card, fills out a form and you're done. Okay. So you get a nice quick hit, but it might only be a thousand dollars or $3,000 a year. And you might start to say, you know what? I want the $10,000

SPEAKER_39: a year customer. I want the $30,000 a year customer. So it really is, um, going to determine how you staff and fund your company. You may need more runway to go after that. So, uh, what does your gut tell

SPEAKER_149: you or what does your knowledge tell you about what type of market you're in? Are there a small number of large flyers and a long tail of small ones? Are there plenty of both? So both business models will work. I don't know enough about trucking to know the answer to that.

SPEAKER_122: Yep. Yep. That's, that's pretty good advice. And right now we're seeing our product into this midsize level, uh, which is the, you know, sales cycle much faster because like you said, the owner is involved. So it's really easy to show them the product. Like, Hey, here's the solution. Uh, they're scaling their operation. So it's really working right now, but then, you know, those enterprises comes and they like, Hey, I want to use your technology, not your product, but your technology to put into our current systems. So, but, uh, but it's a different game, like you just said,

David Friedberg: so we will, um, got to like, you, you might need to price differently for them. And you then have one group that is easy breezy. They'll just use your solution. And like I said, the other group is, Oh my God, I need, you know, this custom thing. I need that custom thing. Can you get into our accounting system? And a way to do that is to say, yes, we can do it. It's going to require a $250,000 activation fee, uh, for us to build that. And we need you to pay up front. If you're willing to do that, we're happy to work on the project. And then you can hire two developers, uh, or a product manager and a developer. And you tell them, I will embed ourselves. You may have heard me talk in the accelerator or founder university. Did you go to founder university too, or just accelerator? No, just the accelerator. Great. Um, and so I've told the story before about bear hugging a client, which is when you embed yourself at their company. So you could say to one of the big ones, I'll put a product manager in your office for six months to sit with your team and to outline all these features. We just need 250 K in advance. Uh, and that's for the first six months of this trial. And then afterwards, this is what the pricing would be. So you're kind of challenging them

SPEAKER_39: to not waste your time. The biggest problem founders get into is that, Oh my God, this is such an incredible lighthouse customer. And then some mid person there wants to get points. They drag you through different specs and documents and meetings, and then they try to sell it to their boss. And then that person decides they're going to leave the company. They get fired and you lost your contact there. So you gotta kind of let them have skin in the game. You gotta say, listen, we're a small company. We're eight people in order to do this. I need two more people in order to do this. You know, I'm adding five new customers a month right now for the existing product.

David Friedberg: I just need you to, uh, you know, make a commitment upfront. And so how do you feel about a 250 K pilot for six months? And the person might say, I can't authorize that and say, Oh, who in the organization can, maybe we should have a meeting with them. And you just challenge them, you challenge them, you challenge them. And if they're not serious, you say, you know what, we were just resource

SPEAKER_169: constrained. We'd love to work with you, but we need to stay focused on this. Can we check in in six months?

SPEAKER_165: And boom. Does that make sense? Yep. A lot of sense.

David Friedberg: Yeah. Thank you. Thank you. Don't be scared of these customers. Look at how many employees they have in times that number by $200,000. So what's the biggest of the customers you've talked to who are the big fish, how many employees do they have? Do you even know? Uh, what a thousand? A thousand. Yeah. So a thousand times 200,000 is $200 million. They're probably spending all in about $200 million running that operation if they have a thousand employees. And if it's a public company, you can go find that out. But you know, it's not that every employee is getting paid 200,000. It's that, you know, maybe on average, the employees cost a hundred thousand with their benefits, but they have to have office space. They have HR, they have turnover. They've got to recruit people. And you start just as a back of the envelope number for a midsize company that, and then for a Google or for a Facebook, maybe it's 500,000 per employee. You can estimate what they spend. Alex, if you were to look up how much money Meta spends a quarter and how many employees

SPEAKER_39: they have ballpark and you were to divide those two numbers, and he'll go do that in real time, because he's awesome at these kinds of things, you'll get a certain expense base per employee. It might be 500, 750K in total spend because they build signature office space. They have other kinds of infrastructure. They have consultants. Free food. Free food, all this stuff. You know what

David Friedberg: the free food bill is at those big companies? It's like 10, 20K per employee. Wow. The free food is so

SPEAKER_39: expensive that the IRS felt the need to investigate if they were end running compensation by using their pre-tax dollars and giving an employee benefit, to which point the IRS said, hey, if you're going to spend $20,000 on Perks for your employees, we need you to pay tax on that because that would be the same as like paying for a corporate apartment, paying for a corporate car that's used personally. Like they have to eat. That's on you to give that perk. And I don't know where that IRS investigations wound up, but

SPEAKER_149: you know, that's how crazy the spending they have per employee is. So I would-

SPEAKER_12: So as far as we can tell, Jason, you were very much close to dead on. 2024, which is the latest full year employee data. They don't drop it every quarter, about $550,000 in expenses per employee

SPEAKER_21: against revenue of $2.22 million per employee last year. And that's why Meta is a good business.

SPEAKER_138: And that's why when they keep the team size the same and grow revenue 10 or 20% per year,

David Friedberg: the earnings increase, which is why people keep betting more and more money on the Mag7. Listen, continued success, Camilo. Hopefully we did a good job for you in the accelerator. Did my team work

SPEAKER_122: hard and support you? Absolutely. Yanka was great. We made a lot of connections, great connections, a lot of business interests. So it was really good, really good. What was the best part of the

David Friedberg: accelerator for you as a founder that you got the most value from? And then after that,

SPEAKER_169: what do you wish we would add to the program that would make it better for you as a founder, if anything? Yeah. Take a second to think about that. Yeah.

SPEAKER_122: The best thing definitely is the pitch sessions, you know, like practicing this pitch, uh, this two minutes, uh, pitch like every single week. So after that, like every time I have a meeting and I'm pitching and they're like, you're done. Like, well, you, you just covered everything because you, you know, we covered the, the issue, what's the solution, his product revenue, and that's it. So that's

SPEAKER_148: definitely really, really. The pitching really helped you. Yeah. And Alex, uh, well, we originally were at like five minutes of pitch, then went to three minutes, then we made it two minutes.

David Friedberg: Two! That's only 120 seconds, man. Yeah. Well, here's the thing. You are making a trailer of your startup is what we tell the founders. Now the trailer for Star Wars doesn't tell you, spoiler alert, that Darth Vader is Luke's father. It just tells you like, this is an epic thing happening in space, and there's lasers and there's ships and it's, there's a Wookie and you're like, oh, I want to know more about this movie. That's what we try to accomplish, which is, oh, wow, this is a

SPEAKER_169: interesting founder or interesting founders. And they found it's a really interesting market. Oh, and this is a really interesting solution. Oh, and this is a really interesting Tam. Oh, and the

David Friedberg: competitive set is, you know, Marge smoking those Marlboro lights, two packs a day, you know, yelling into a landline, uh, and who tells you to get the hell out of her office and throws the ashtray at you when it's filled with 20 cigarette butts. If the boss comes in there and tries to give

SPEAKER_198: her a hard time. Two packs a day is excessive smoking. If you're smoking two packs a day, man,

SPEAKER_154: what's inside of you? I mean, I, I know these dispatchers. They're there. These are curmudgeonly hardcore individuals. I love them. I knew them from the dispatch at, uh, Harbor

SPEAKER_149: car service in Bay Ridge, Brooklyn, where I grew up. Uh, cause yeah, there, there was like a large

SPEAKER_39: marge there too, who would yell and scream at people and get the hell over there. Pick up John to be its kids. It's the middle one, Jason. He's, he's gotta get to school. He's late. Uh, they give him a hard time. Okay. What can we do better? Is there anything that you would have liked

David Friedberg: to add to it? Sometimes people ask for mentors. I always felt the mentor thing was kind of garbage

SPEAKER_37: because they had these mid mentors, but you, you tell me, is there something that we could have added to the program that would have helped you more? I mean, well, yeah, the thing with mentors

SPEAKER_122: is the, that really depends, like who's the one that is going to help us, you know, uh, if it is going to be like a, you know, founder, maybe, uh, uh, could be a good advice. Um, but what else

SPEAKER_207: I can say, uh, more money, zero, we'll double it. Sure. Great. Yeah. And more money to the program.

SPEAKER_122: Thanks Alex. You just cost me more money. That, that would be a, a great opportunity to double the money. Yep. And, um, but not for real. It's a really full program. Uh, we really love the, the, the pitches, the introductions. I mean, we haven't seen, um, well, programs like this one where you see every single week, five, six, seven different VCs and just not like associates, but also like, you know, the managing partner, and then he's the owner of the, of the fund.

SPEAKER_120: So that, that adds a lot of value. To unpack that, Alex, you know, some firms do like this big demo day and they advise you, Hey, don't do meetings with investors before then just get

SPEAKER_138: your startup nice and tight. I took a different approach. I'm not saying the other approach doesn't work. It does to create like a bit of a marketplace at the end. What I'd said was, let's pick our favorite investors who we have deep relationships with, who give great advice and who are willing to give up an hour and a half of their time. And then we say to those ones we like

David Friedberg: best, here's the calendar, pick a date you want to come. And then it becomes competitive. They're like, I want to get these companies in the first, second, third week because I want to pick them off before other people see them. And so then what happens is, you know, Camillo will get some great piece of advice, you know, three or four times where people are like, Hey, this Tam is you're charging too little for the value of creating. And when a founder hears that not from us, not from their customers, but from three or four VCs in three or four weeks, and then they just try raising their prices and then they present, Oh yeah, we have three pricing models. There's the introduction for, you know, one truck company or two, three and under companies. And we charge them, you know, a hundred bucks a month. But then these ones we charge 750 a month and 1500 a month. And when we changed that pricing, we saw people, uh, half our customers moved up and we deprecated the free product. We got rid of 20% of our annoying customers and the 80% said no problem. And they just, you know, agreed and checked the box to pay the higher rate. So we record every piece of feedback, Alex, we normalize it and we say, okay, you keep getting this question about your pricing. Should we have a discussion about that? And do you want to read a module we've written on pricing? Or would you like to talk to an expert about pricing? And I find this model works better with great founders where we're not dictating to them how to run their company. They're just collecting lots of feedback, lots of feedback. And then they themselves normalize it and make their own decisions and we help them by normalizing it and taking out all that work. If you had to set up those seven meetings each week for 12 weeks, it would be a full-time job to get half of that done with the associates at the firm. There's nothing wrong with associates. We have plenty of them, but you know, like I said, we try to get more senior folks. The associates though, I will say can be

SPEAKER_39: sneaky, influential in some firms where you have lazy VCs who are like, they just send all the associates. I can't meet with all these companies. We do a hundred meetings per week at our firm,

SPEAKER_149: introductory meetings. I could never do more than 20 of those, you know, back in the day, I could do a hundred, but man, that becomes like hard. Your brain will melt. All right. Listen, great job. Very excited for you. Thank you for letting us invest. And thanks for the kind words

SPEAKER_122: about the program. All right. Thank you guys. Appreciate it for the time.

SPEAKER_17: All right. Next up on the show, we have Elise Myrens. She is the CEO and co-founder of 10X. It's 10X AI.

SPEAKER_12: Jason, this is a company that wants to get more information about potential risks, help home owners, homeowners, harden their houses to prevent wildfire damage, and also ensure that insurance companies can properly underwrite potential risk. It's an enormously important business because we've seen more natural disasters, higher repair costs. The entire U S insurance market seems to be a little bit upside down. So I'm very excited to hear more from her and here she is. Elise, how are you doing?

SPEAKER_220: Hi, I'm great. Thank you. How are you?

SPEAKER_09: Nice to see you again. You are from the launch accelerator 35th class, if I'm remembering correctly. 34. 34. Okay. So I'm not remembering correctly, but I was close. I was in the ballpark. And we had the chance to partner with you on your new company. Maybe you could tell us a little bit

SPEAKER_02: about what you're building and why it's important. Yeah, happy to. So 10X AI protects homes and their insurability from extreme weather. So I'm talking about things like flood, wildfire storms. And essentially what we have built is software that uses computer vision and AI to assess risk at the individual property level. So at a home level. And what that does is it gives insurers a granular view that helps them understand the true risk of that property, which enables them to underwrite and ensure that property more accurately. And what it also does is it enables the homeowner. So the people who live there to understand their exposure to risk and to take actions that can actually reduce that

Chamath Palihapitiya: risk as well. So AI is the better way to do this. How is it currently done? And why is AI a better

SPEAKER_09: solution? And maybe you can show us because you had some really interesting demos and in your deck and when you, I remember when you presented to Sequoia, uh, I was there and they were fascinated by your

SPEAKER_02: approach. Yeah. Yeah, absolutely. So to date, what has been done is insurers would normally make a regional, uh, sweeping decision based on regional models about the exposure of a property. So they might say this entire zip code or neighborhood is this general level of exposure. And then they use that to inform how exposed that home is to a particular peril, such as wildfire or flood or hail or winds. Now there's the introduction of some manual inspections where a person or a professional would be booked and organized and they would go to the property on behalf of the insurer and they would do an inspection that could take a couple hours. They got to write up a report, ingest all of their written and verbal notes to put together a report to explain that exposure and then hand it over to the insurer who then has to adjust it into their underwriting process. What we do is about 80% more efficient than, than that process because we have built, as I mentioned, AI and computer vision models that essentially enable images which can be collected in up to 30 minutes max on a mobile phone. So we can just take their cell phone, walk around the property with our app. Very, very simply, you need zero expertise whatsoever. The average homeowner can do it. And we can upload those images and run it through our proprietary models to be able to diagnose the exposure of that, that property. The other thing we're doing that is quite unique is we are now starting to send drones to properties as well. And this is really important because in some cases insurers would actually prefer that the homeowner not be the one collecting the data. It just offers another level of verifiability. Sure. Yeah. As we've seen,

SPEAKER_09: homeowners can value their homes and maybe get into shenanigans on the margins. And this is very interesting to me, this company specifically, Alex, because when we had the tragedy in the Pacific Palisades, you know, I'm just playing a quick video here on YouTube without the sound. I'm assuming you two can see it. Yeah. Yes. So this home was made to be fire resistant. It has a number of features. And if I were to zoom in on this for a second, let me pause it there. Perfect. Now, when you look at this home, what you'll notice is it doesn't have overhangs of the roof. You'll also notice it's fire retarded material everywhere. And around the home is a giant stone fence with landscaping with

SPEAKER_37: a lot of gravel and metal and not a bunch of trees around the home. You can build homes today that in a wildfire would be resistant to them. But this home, Elise, would have been bundled with

SPEAKER_237: all homes in Pacific Palisades. Correct? Spot on. And that's because of the way that risk has been

SPEAKER_02: being calculated by insurers, I said, in these sweeping regional models. And so what happens is even if your home was lower risk, so in that case, that home's insurance actually is probably higher

SPEAKER_241: because they're technically compensating or paying for their neighbors or regional risk. Right?

SPEAKER_12: Now, Elise, if I understand, this is not because the insurance companies are lazy, but because there's actually a regulatory barrier they can't cross or couldn't cross until recently.

SPEAKER_241: Yeah. There's a couple of crazy forces at work, which are kind of unexpected. And so one of them is

SPEAKER_02: actually in California. Until about a month ago, insurers literally did not have regulatory-approved catastrophe models that they were allowed to use that were forward-looking that enabled them to calculate the exposure to true and up-to-date wildfire risk, which is bananas. You would think that that would be very obvious, but it hasn't been the case. And so as a result, of course, they would be pulling out. They literally haven't been able to measure risk accurately. And so there are now three catastrophe models that have just been approved in the last about four weeks, Verisk and Moody's among them. And what is really cool about these, if you are nerd like me and get into the weeds and looking at these models, is that they have placeholders within them where you can insert data about specific secondary factors related to the home. So we're talking about characteristics, like what material is the fence made of? Is it combustible? And does it link up to a large patch of vegetation? Because you may think you have defensible space around your home, but say you have a wooden fence and it connects, you literally have a wick to a combustible source. Or what type of windows do you have? Are your vents covered with the right mesh? And what's really cool and what we factor in mathematically to our algorithm so we can actually tell a homeowner and tell an insurer and tell a community how exposed they are, is we have risk factors with each of those characteristics of a home. So if you have the right type of roof, you're reducing your risk by six times. If you have the right porch material, you could be reducing it by three times. If you have vents that are covered by the right size mesh, so embers can't fly in, you could be reducing it by 1.5 times. And it goes on and on. And a lot of these things are not things that people think about every day. And so that's what we're able to offer people really simply.

SPEAKER_12: This goes to the question that I really had for you, which is who is the customer here? Because I can imagine the insurance company being the customer, or it might be the individual who wants to get a better rate, so they might pay for what you guys put on. So who do you sell to in the market?

SPEAKER_237: Yeah. So insurance carriers are our ICP. So those insurers who have exposure to perils like flood,

SPEAKER_02: wind, hail, storm, wildfire, which increasingly is kind of everywhere, as we all know, for anyone picks up the news, the homeowners themselves are the users. And so they would be prompted by their insurance provider, either if they are at a point where they're looking for a new coverage. So if an insurer is deciding how to price the risk, and they need this kind of data, so that a homeowner could actually get rewarded with discounts, which is now there's new regulation that helps enforce that in California, for example. And then they would be able to access that software,

SPEAKER_248: basically through their insurer. Yeah, this is, I am obsessed with this. And

SPEAKER_09: because we had a, we had an investment. It didn't work out in a company called Blockable. They were making prefab homes. Now you might think, oh, a prefab home, oh, that's just going to burn up. No, au contraire, mon frere. What they were doing was the gaps in these homes and the materials that you can build in a factory, as opposed to onsite, because cutting certain types of materials is, cutting certain types of materials is impossible in the field, Alex. And so Dwell did an interesting series on this. And especially in Australia, this is a big deal because of the wildfires in Australia are, I think they make some of the wildfires we have look like, you know, campfires, not to make light of anything here. But as you can see here, this is a fire risk infrastructure. This is a Dwell story, great magazine, dwell.com. And you can see they're

SPEAKER_37: using steel for the roof, et cetera. But at least the gaps is a key because if there are tiny, even the tiniest of gaps, the air gets sucked in. You have all kinds of suction kind of situations that happen

SPEAKER_09: with heat that you've all experienced when you open an oven, et cetera. And what that does is suck embers into the home or it sucks embers into, you know, a pile of leaves in the gutter, all this kind of stuff contributes. And they have been working on this specifically to do fire resistant timber. Here's an example of a timber that is fire resistance. And then other folks are making what some people might consider ugly, you know, made out of iron, but you don't have to go that far. And then as you see here, the distance of the trees to the home also critical. So it is completely possible that if you clear the brush around your home, the chances of it burning down go way up. If you look at flood resistance home, that was another thing that happened in the wake of Katrina and Breezy Point where I used to summer as a kid in Brooklyn and the Rockaways, they started building homes that are four feet above the water. Yes, at least. And have these really interesting panels on the side of them. My brother was a firefighter, my grandfather as well. And what these panels do is when the water comes rushing in, you know, so you don't see under the house, when the water comes rushing in, it just blows the panels out. The pressure from the water just blows them out. And then you just have, you know, whatever you have in your basement is like, you know, you keep surfboards there or whatever things, you know, you're not putting a full on, you know, theater in your basement. These panels blow out, allowing the water to flow through. So aesthetically, it doesn't look like you're living on stilts, but essentially you are because the panels just are made to break away. So flooding is the other big high order bit

SPEAKER_244: here, Elise, I guess. Yes, flooding is going to be second for us. And I think you're touching on

SPEAKER_02: something, Jason, that's really important. And we hear this so often when we talk to homeowners who say, well, what can I really do? I live in an area that, you know, I'm told is prone to, in this case, wildfires. The reality is there are so many things that people can do, but they don't know. And it's not only a roof, you know, and then people, you know, say it might be the roof. They're like, well, that's so expensive. Well, the reality is the numbers have been run. And we know without a shot of a doubt that at the very minimum, $1 invested in the right retrofits or controls on a property saves at least at a minimum six in recovery cost, right? And then if you can do these types of retrofits and then get insurance as well, so you're insurable and you're not having to pay for, you know, an insurer of last resort that is insanely expensive, well, then you're in a really good position. And so we're trying to give people advice on those things that they can do to help

SPEAKER_253: maintain their insurability. Elise, I have a kind of annoying question here, but if we make it so that

SPEAKER_12: people who are very proactive about hardening their house from either fire or flooding or whatever, uh, can be insured on a one, a single house basis. And is that going to leave a lot of homeowners who may not be able to harden their homes out in the cold, essentially, because they'll become uninsurable because the risk pool will be narrowed from a neighborhood size to an individual home? What's the other side of the good news that we can have more information, I guess?

SPEAKER_244: Yeah. Okay. So there's a few aspects of that. One is actually something that we are seeing as interest

SPEAKER_02: from communities themselves. Like, so from HOAs, local and regional governments who actually want to take a community-wide approach to risk reduction. And so the idea here is that, you know, when you can harden multiple homes in a neighborhood, you can actually improve the survivability of the entire neighborhood. And we also see from a behavioral lens, when one person in a neighborhood starts to take action, it does spread. I was going to say like wildfire, it's a wrong pun. Um, but it does,

SPEAKER_255: it, there is a contagion, you know, in a certain aspect of, these are terrible, terrible analogies.

SPEAKER_265: Terrible analogies. It smells like a smile. I could have used, that would have been even worse, Alex. But in some ways, those homes at least become a firewall. And there would be an argument in my mind,

SPEAKER_37: Alex, to, um, even make the disparity between the safe homes and the unsafe ones even greater. To incentivize. To incentivize it. Because if, let's say we lived in a, you know, a thousand home plan, you know, planned community, that entire community could save the three communities next to it. When the embers start flying, they, they just are natural blockage. And you might look at it in a community and, you know, the, um, the, the floral arrangements and the, and the, and the parks and everything, uh, and the structures that are not homes.

SPEAKER_09: You know, those are kind of a tragedy of the commons, like who's going to change those. But you could say, you know what, these trees are no bueno. There, there are some trees in California that are literally like, I mean, you, you, you could put dura flame logs around your house and it'd be more fire resistant than some of these, like the, are they the, the palm fronds?

SPEAKER_265: That the, one of the things that fall off of the palm trees, these giant things, I don't know what

SPEAKER_241: they're called. Yeah. Palm somethings. Anyway, those light up like tinder, tinder. Yeah. Oh my God.

SPEAKER_265: And then they break off and they fly like a, like a, literally like a missile and land on somebody's

SPEAKER_09: roof. And then that person's got, you know, a shake roof and okay, game on now. So yeah. So, uh, we're here for a little office hours. Um, how's the company going? Is there anything we

SPEAKER_278: can help with? I love that question. Thank you. Yeah. Things are going well. Um, we're working

SPEAKER_02: with a handful of insurers that are helping feed into our product development, which is very exciting. Um, my question for, well, there's a couple of things. One is actually an offer that I'm going to make, which is for, um, any insurers, brokers, um, and even homeowners who might be listening to this, um, who are really interested in figuring out how to protect their home and understand their, their risk. Um, we're currently working with some partners in California and essentially if there are people who would like to, um, get an early version, uh, test of their property, contact me at Elise at 10xai.com. That's at T-N-A-X-A-I.com. And I'd be happy to, um,

SPEAKER_00: put you in the queue for some of that. Great. So we need some, uh, vanguard here. We need some early adopters who want to get early access. So great. Let's get some of those going.

SPEAKER_01: And then the question I would put, um, this is something I'm loving brainstorming around right now,

SPEAKER_02: are different ways to hack sales cycles and really regulated industries that are slow moving, right? So in that B2B space, I love talking about this and I'd love, I don't think you and I have talked about it. I would love to, to hear your thoughts on that. You know, there's a tough one,

SPEAKER_09: you know, whenever you go into education, let's say, um, or you go into housing, uh, as you are, um, healthcare, these are some of the giant castles that are hard to scale. We've got moats around them and they got drawbridge, it's got alligators in there, got arrow slits. It's not fun, um, to scale those. One thing I've seen is companies like brilliant.org, um, and Alex will pull it up. Um, you know, they thought, Hey, we can help educators, right? Um, teach kids math more efficiently, et cetera. But I think one of the lessons Sue learned over there very early was

SPEAKER_37: we can actually, uh, sell into the top parents, um, and have those parents, um, you know, use these tools to make their kids superhuman to use a term, uh, another startup that we're investors in. Um, and then those people bring it to their schools or it becomes, you know, oh, the kids who are doing really good in the class are using brilliant.org and their STEM stuff. And then they say, Hey,

SPEAKER_09: can we get this for our school? And so I wonder if you find high-end homeowners where you can consult with them or build some product or service that helps them navigate lowering their insurance and, um, also just lowering their risk and advising them. So if there was an audit of my home and you told me for $500, a thousand dollars, $2,000, $5,000, you know, I have some expensive homes, uh, that costs a lot of money. And somebody said, Hey, we can audit your home for $500 and give you recommendations right now of how to lower your insurance, or we can go call your insurance company and tell them you've done these things. And whatever we lower your insurance by, we'll take half of, uh, the gain, you know, there could be something there. Now, I don't know enough about the insurance business and lowering insurance based on these remediations, but that could create, uh, a way for you to integrate yourself into the industry, to just raise awareness. And I, I think I spoke to you a little bit about

SPEAKER_37: you're so passionate and you're so great at communicating this and you have that just natural enthusiasm, a YouTube channel, just about this issue where you have homeowners, you have builders, you have the person who makes the roofs. You have the person who is a contractor who you met in Santa Rosa, you know, uh, up by Napa where they've had some terrible fires and does this kind of retrofitting and, and just short videos, even tick tocks. Hey, my name's Elise. I'm from 10 X AI. We help fortify your home from floods and fire and lower your insurance today on the program. I'm going to show you this tip or here's a tip for, you know, uh, stopping wildfires and dealing with floods.

SPEAKER_09: Those kinds of things could just make you, you know, the queen of this topic and that would naturally start you off, uh, on second or third base with partners. I know this because people, when I had this week in startups and I started my angel syndicate on angel list and then moved it to the syndicate.com We're like, oh yeah, I've, I watched this week in startups. I'd love to be part of your syndicate. Or I saw a 10 X AI. Are you going to syndicate that company? So we inserted ourselves into the discussion and then we became the host of the discussion. We became the place where the discussion about startups was happening. The discussion about angel investing was happening, which is why I wrote the book on angel investing, literally called angel, uh, you know, now in 12 languages. And so I had that calling card that just made, you know, any meeting request, uh, happened very quickly. And now with all in, it kind of moved over into pop culture. So when I go to another city now, and I want to talk to people about bringing founder university to their region, or, you know, I want to record an episode of this week in startups with a founder. We just had a founder on the pod from neuro and he was referencing Alex, a nice conversation from previous episodes about self-driving. And so that means their chance of coming on the pod goes up. The chance of me investing in the company goes up. You get the idea. Uh, all right, uh, everybody go to 10xai.com t-e-n-a-x-a-i-t-e-n-a-x-a-i.com. And if you are a homeowner or in your insurance and you want to help a founder out just with some conversations and, uh, maybe, uh, some tips and advice, at 10xai.com e-l-y-s-e at great job. Thank you for letting us invest in your company and we will see you all next time on this week in startups.