Roofing is About to Change Forever | The Roofing Insights & RBP Podcast

Give me one AI prediction for roofing that people will laugh at today but believe in 3 years.

Adam Sand will have a humanoid robot on a roof nailing a shingle in 3 years.

Because the home buyer group is older, you are going to be able to get away with using crappy AI longer.

AI is like plastic surgery, you only notice it when it's bad.

There are people using AI in their marketing right now and it's allowing them to build more good stuff. You wouldn't even notice.

Good marketing and bad marketing is not AI dependent.

In the front end, you have You have to prove to AI that you are number one choice, and you're not number one choice because your ChatGPT drops 5 Facebook posts today.

So you're saying to all the business owners, don't use AI for marketing sales, don't use it for your front end, but you're saying on the front end you need to be human.

Even as a big pundit of AI, I am investing more heavily in my own personal brand today than before because yeah, the only thing that matters is me talking to that camera and people hearing real words out of my mouth. AI is like plastic surgery, you only notice it when it's bad. And so yes, people are using AI, and yes, if you see the em dashes and the word Delve, it's a pretty safe bet that it was written by AI. There is a line, just like plastic surgery being bad. Where's the line and how do you make it good? So what is the purpose of marketing? It's to invoke an emotional reaction that makes your brand memorable so that people know that you exist and might want to take interest in your stuff. That's the goal. Good marketing and bad marketing is not AI dependent.

To your point, Adam, you're saying like there are people using AI in their marketing right now and it's allowing them to build more good stuff. You wouldn't even notice because they're following up with the right— it's the people that are using as the lazy way out, aren't putting the effort into it, aren't double-checking it, that are—

yeah, I got a Christmas light contractor who tripled his business last year, and he did these really funny Christmas light ads where a guy was like tied up in Christmas lights because of the whole tangling of Christmas lights problem. Like, don't be held hostage by your Christmas lights. And then he's also built something like Dope Marketing but in Canada, where he takes like 1,000 houses and puts Christmas lights— so changes the Google Street Image to like dark and then adds Christmas lights and makes them their own personalized postcard. And it's all automated by AI and then delivers them a postcard saying like, this is what your house could look like this Christmas. And he's getting like returns on an ad spend that are unheard of. And Part of it is because it's effective. The other part of it is because it's ridiculously low cost. It's costing him 7 cents. And that's why I like the saying that AI is like plastic surgery. You only notice it when it's bad.

Well, let's talk about the demographic here too, right? Like my parents are probably going to have a harder time identifying AI than my younger siblings, right? Like the older crowd, no offense to them, but yeah, younger people spot it easier, right? But like in our industry specifically, who is the target audience? Boomers own the majority of houses, right? Right?

Like it's typically older clients, but it's also changing. Like, you know, your customers also getting younger and younger.

I don't think the average homebuyer age has gone less. Someone check that. But yeah, I think it's like been increasing or maybe it's coming down a little bit because the homebuyer group is older. You are going to be able to get away with using crappy AI longer, but you're going to fall behind if you're selling tech products and stuff to a younger audience, makeup, whatever else you might want to do to a younger audience. You're going to have a much harder time pushing AI slop and crappy AI. They're going to notice. I think we're going to go to a point where yes, our industry is going to fall behind because people are going to be able to get away with crap.

So this is how I see AI and user experience aging. It's dropping off the cliff. We're getting used to this stuff so fast. You cannot even compare what we did a year ago. 6 months from now, we are so used to AI, AI post and stuff that we're just scroll right through it. And right now, Instagram, Facebook, all the platforms, they recognize it. So here's what trending today, love it or hate it, but polished content, studio content like this, it's actually old school. Like nobody's doing it anymore. Like platform wants you to be not only original, but they want you to be eating burger and talking to the camera. I mean, can it be AI? Yes, but it's also not polished, raw content. Here's my take on AI-generated stuff. I see companies out there who trying to streamline You know, the message is don't do anything. You don't have time for it. Plug it in here and every day is going to put— like, imagine following business account. I don't care how big you are, Ford, Tesla, or a local roofer, that every single day it's AI-generated copy and image versus somewhere where a business owner like Gary Vee says, if you're the business owner, you can go turn on your camera every day, say, my name is Dmitry, I'm a founder of directory.com. Here's what I do. Here's the problem I solve. And every single day you're going to find more audience and more, even if if it's the same message, you still gonna attract some people because it's you sharing what you do. And compare it to business who post daily, but it's all AI.

One of the highest engagement posts we got was 'cause AI messed up a photo, which led to people talking smack in the comments, which then led to like a ton of people, it was the most engagement I think we've ever got on a post on LinkedIn specifically, it was like 60,000 impressions. And people like instantly were like, oh, no one else knows. Then they get to the comments and there's all these other discussions. So like if you're gonna use AI, just use it really, really bad. To the point that it offends people, and then you get engagement too.

You're speaking with the concept of like AI commoditizing a race to the bottom. We saw this in text, right? Like now nobody— it's really hard to find a really great writer that you wanna follow because you see a long piece of text and you just assume it's AI, even if it doesn't have all the AI tells. There's a thing called Humanizer. So anytime my AI writes something, there's a skill that references a Wikipedia article where everybody maintains this Wikipedia article of all the tells of AI writing. And then anything that gets written by my agent gets humanized. Basically, it references that Wikipedia article and removes it, right? And that gives itself a score and it can't write, it can't push anything out unless it scores a 75. That's something that had to come in to quality check AI writing. The same thing happens in video, but you're talking about the commoditization effect where like faceless YouTube videos suck and even like AI-generated face videos suck. And so you're right. That the real commodity, the real value going forward— like Dimitri's YouTube and personal brand was like 6 out of 10 valuable 5 years ago. Today it's 11 out of 10 valuable because people want real people, because trust is going to become a moat going forward. So everybody should absolutely be investing in their personal brand and the trust that they've built. The one weakness in Spencer Pratt's campaign is that when he was able to generate as much attention as he did with his funny AI videos, there just wasn't enough content out there of him personally speaking on his issues because he was so new. If you attract and build attention and get awareness of your thing out with something that is good AI, right, the not bad plastic surgery but the good plastic surgery, you get the attention. Then the goal is to then bring it back to say, oh, is this trust? And then there, there's real content. But there's different types of products, right? Like if you're marketing skinny tea in a water bottle, the days of paying a pretty girl $5,000 for 3 posts about your skinny tea, when you compare that to $16 in Higgsfield AI credits, there's a value change there where people don't need as much trust to know about a new water bottle with a different handle or a skinny tea that smells like lavender. They don't need as much trust for that. But when you're talking about the level of stuff that you do, Sam, or what I do, or Dmitry, what you do, your personal brand attached to directory is worth way, way, way more today than it was before 'cause directory can be replicated. The technology, the website, even getting the network out there can be replicated. But the other competitors to directory don't have the trust of a human person with a reputation. That's what you've got. That's that moat that you're talking about. And in that lens, completely agree with you. That's why I'm sitting in a studio today. Even as a big pundit of AI, I am investing more heavily in my own personal brand today than before because, yeah, the only thing that matters is me talking to that camera and people hearing real words out of my mouth.

Adam is done talking. There's nothing to add to it.

Yes, I agree with Adam is that personal brand matters. Humans like humans. I was on the phone with someone earlier and they were talking about how they spent hundreds of thousands. They own a roofing company in Minnesota. They've spent hundreds of thousand dollars on marketing. And I said in the conversation, door-to-door sales beats everything every time, no matter what, no matter what you spend on marketing. Door-to-door sales is the best return because marketing, you're throwing the money in, you're hoping to get it back. Door-to-door sales, you're typically not paying a rep unless they bring someone in. Like, but I genuinely believe it's because there is that human interaction and humans like humans, no matter how much, like even like I sold all through COVID and people were like hiding in their houses, didn't want to talk to anyone. Like, yes, human interaction outsells everything when you're selling to humans. Adam, the question I have for you next, which is going to be a long answer, I know it's coming. What happens when consumer AI gets so good that it starts blocking, right? Because we had— imagine on the internet we had all these ads and pop-up ads and then people got ad blockers. Right. Or like, oh, well, people start spamming texts. Well, now you got, you know, your iPhone automatically detects or someone tries calling you. Now it's a spam alert, right? What happens when people's technology starts blocking all this AI?

I don't think it's blocking AI. It's that the consumer is catching up to the enterprise in terms of AI use for consuming. And tools are being made. If you watch the Google I/O event, they talk about all the consumer technology that's coming in. Good example. You see a blue dress. The manufacturer has put all the specifications of that blue dress online. So now when their AI knows your body composition, you could do a virtual try-on. Well, this blue dress is available from 5 different retailers. So it's like Trivago for travel, right? And you say, I like this blue dress, it's priced at $70, but I'm a buyer at $50. So you can put it on a Price Shop watch list for $50. Well, then Google's going to go and sell and say, hey, all the people who sell this blue dress, We know, we think that you could get 1,000 buyers if you reduced your price to $55. Distributor, a seller can basically put out there that they're, they're price dropping to $55 and then everybody gets that notification and buys the dress. So that's one way where like it's literally agents talking to agents, like the revenue agent at the Blue Dress Company is telling them that there's a market of buyers at this price. And then there's a bunch of humans that have said, I'm in the market at this price. So that's one piece. Another piece that really needs to come out, just like when we had credit cards going into computers, people didn't trust it. People don't trust AI agents to buy yet, but agent-to-agent commerce is already out there and Stripe is already building for it. And if you go to isitagentready.com, you can like test your website if it's agent ready, and then you can start to work towards making your website agent commerce. And for perspective, Everybody here is old enough to remember when they had a business and people said your website needs to be mobile responsive because 20, 30, 50% of website traffic was mobile. Now, do you remember that phase? Oh yeah, right. The people who responded first won. If your website didn't work on mobile, people just went away. People started losing business when traffic was at 30, 40, 50%. 2 months ago it was at 51%. That's when that Is It Agent Ready website was launched, was when it was at 51%. Yesterday It was reported that 57% of all website traffic is agents now. So the consumers are not going to block AI slop. They're going to shop through their ChatGPT, through their Claude, through their Gemini, through their Google app. And, and Apple's recent event is all about trust and safety, and buried under all the stuff for kids is the stuff about agentic commerce.

Okay, Dmitry, I have a question for you. Given all that, because Adam's pretty much saying that like commoditized services, commoditized products are neutralized. There's no winning anymore. Like, it's not that you weren't that aggressive with it, but like, essentially, like, you're not going to be able to use AI to get through this stuff because they're going to be shopping through the AIs. How do you— if you own a roofing company or in the roofing space or home service space, right, because that's the majority of our audience— what can you do to prevent your company from becoming commoditized? I know we've talked about before, but if you had to give like the one thing that you need to do to stop looking like every other roofing company or to provide value beyond, what do you do?

I'm not actually buying AI. I'll give you two experiences from this week. Experience 1, I have a broken car and I was looking for mobile mechanic. So I went to ChatGPT, find me local mechanic. ChatGPT gave me like 4. I called all 4 of them. I could not talk to a person. Nobody is returning, nobody, right? So I found mechanics, but I did not find anybody who can actually help. So that one experience. Second experience, I bought furniture from Home Depot and I bought one of those assembly $300 to put it together. So furniture arrives on Wednesday and I'm texting actually the number like, hey, when is assembly? It's supposed to be Friday. And now they saying it's like next Tuesday. So I have like 16-year-old who can assemble it. I'm like, can I cancel it because I don't like this experience? So I'm chatting with a chatbot. Same thing. I'm not getting my answer. I paid $300. There's no human on the other side. I'm frustrated. So yeah, I have all the tools, all AIs, but I don't have a human on the other side. To answer your question, You have a human to answer. You call me, I answer. You come to me. That's— I'm changing my business model to that concierge service directory. You— I don't want to just give you like, oh, here's your 3 contractors. I want to be like, hey, Mr. and Mrs. Smith, you in Chicago, Illinois? I have 3 guys next to you. Like Joe Bridge, like he's amazing guy. Like, right? People want interaction. People do business with people. Build your personality, build your expertise, become number 1 in your space. I don't care if you're pool guy, if you're a mechanic. When people need your service, they find you. If they trust that you're number one expert in that zip code and whatever, I feel like AI later will be able to say, hey, this is the guy who is number one in this area.

So you're saying to all the business owners, don't use AI for marketing sales, don't use it for your front end. You can use it, you know, are you okay with them automating for finances, invoicing, maybe stuff during the process, backend follow-up? But you're saying on the front end, you need to be human.

On the front end, you have to prove to AI that you are number one choice. And you're not number one choice because you ChatGPT Drops 5 Facebook posts today. That happens when people around you vote for you, review you, give testimonial about you. When you truly number one in that space, that's it, like you made it. But that does not happen on advertising. It happens when everybody's talking about you, right? Like when people actually have experience. I would argue that the referral is number one marketing, not door knocking, right? Because for every customer who hired you, that's a referral. That's the people who's talking about you. I mean, you can door knock 100 doors a day, but when you do good job, that's where you another good job from that good job, and it just multiplies.

Referrals are the best ROI.

What we're really talking about here is products versus services, first of all, right? These things exist on a continuum. So it's like a deliverable versus an outcome. To Dmitry's point, absolutely, the, they ask, you answer, experience, expertise, authority, trust, like the signals, right? It comes down to like one-way door decisions and two-way door decisions. So like, you know, getting a dress is kind of a two-way door decision. You can return it. Then there's like risk. You will look at 4.8 stars and 5 stars for a nail salon within 15 minutes of you, But if one you can get in today and one you can't get in for 2 days, you'll pick the 4.8 to get your nails done today because there's a level of risk that's quite low, right? Whereas roofing, you might make a decision between 4.7 and 5 and then availability, right? And so what you're really seeing in your experiences, Dmitry, is exactly the problem that I was talking about, which is that the consumer in your case, you were willing to experiment with being an agentic commerce customer. But the businesses in your market have not become agentic commerce businesses. They have not fulfilled that because it's hard. It's, it's hard on the product side. It's even harder on the service side because if you want to sell a service, there needs to be some of that business primitive, that knowledge of the problem. A customer, when they're phoning a roofing company and they need to feel like they moved meaningfully towards solving the problem, like when you first start the roofing company, Dimitri, you got an Angie lead way back in the day when Angie's List was good. You pick up the phone to the lead at 7:00 at night and the customer says they have a leak in their garage and you talk to them about like, where is it coming in? Is it coming down a beam? And they're like, yeah, like, oh, it could be. Do you have like a second story right above, like a bedroom? And they're like, yeah, it could be the wall flashing, could be going behind there. And they go, you know what? That makes sense. And now they feel like they moved meaningfully towards solving the problem. And then your close rate was 70%. So then you're like, I got a business. So then you hire a receptionist and a sales guy and the close rate goes to 25, 30%. Why? Because the receptionist wasn't moving the customer meaningfully towards solving the problem. And there was no prebuilt relationship when they got to the house. We're solving that today with things like inside-outside sales. But that is really just a mechanism for trying to get the deliverable, which is the customer moving meaningfully towards solving their problem. On a service business with a one-way door decision that is high risk.

I want to quickly touch on what can a roofing company do today with AI that was flat out impossible 12 months ago? Like, what's the biggest change, Adam? You're on the front line.

Yeah. So I was just asked this question today by John Starry because I reached out to a bunch of our Roof Claw customers and I said, hey, how's your Roof Claw? Like, anything cool you've been getting it to work on? And he listed off a bunch of things, but he said, my LSA agent has been the absolute best implementation I've done of this so far. So we released an open source, anybody can get it, it's free. They can just plug it into their agent and it will allow them to monitor Google LSA for replies. It'll get that initial reply out. A lot of customers on Google Local Service Ads and any roofer in the room will hear this. Sometimes they don't give you the phone number and you have to click that little button that says click to reveal phone number. Sometimes they don't give you the email and then they have like a short question. Sometimes it's a guy saying, I got 30 bundles of shingles, how much do you want to pay for them? And other times it's a legitimate request. The whole name of the game is response speed, right? The velocity and how fast you respond and the frequency. So this thing will actually open up the browser when you get the email, go into Google Local Service Ads. First say like, is this some guy trying to sell us 30 bundles of shingles or not? If there's a phone number, click the thing to show the phone number, scrape all that stuff, put it in my CRM, but get the conversation going right away. And then that way a human has time to like get in there. If they're an inside sales, it's doing a job follow-up call and they're not able to respond to the LSA lead right away. So that's been his biggest thing. The other thing that was even more interesting was another commercial roofer we have. One of the hardest parts about commercial roofing is that there's a ton of work and they go to these competitive bid boards and there's tons of jobs on there. And then they'll put a document on there that's like 700 pages long. And Section 7, there's 49 pages about the roof spec. And so you're stuck downloading these huge PDFs, trying to figure out which jobs you want to bid. So his Roof Claw would go to the bid boards and it scrapes all the bids, looks for the ones that meet his criteria. It doesn't require a bid bond. It's not close to a water source, so it doesn't need that insurance for water run. All these little things, right? And it puts in like the daily bid brief package. And then he's even got it now where it can do the scope. From a very like choreographed inspection, like clockwise around the building, walk the roof like a lawnmower. Like everything is very specific. The third picture is always the—

walk the roof like a lawnmower.

Yeah. Like mow your lawn. So it knows that the 17th photo is like in an order. So it knows this Canadian thing. Yeah. Right.

Yeah.

We don't do the Hatchi thing that you guys do. We don't have that kind of money up here.

You don't use fricking underlayment in Canada. Okay. How many—

only on new construction.

Have you guys tried Democracy yet? No, sorry, we're working on—

we get internet next week. The concept is like he's taken all of the hard work of like finding the right projects to bid and put it into the point where like every day his estimators are getting the daily bid brief and it's like 5 projects to bid from the 30 that are on this list. And then the second piece is actually writing professional scopes, cuz he's got these guys who are really good at doing inspections and really like understand what's wrong with the roof, but they suck at putting it into words. And so we put all the NRCA docs in there and everything like that. And so this agent can take the inspection and actually write a really good proper scope. He said like a job that took me 2 weeks, I did in 2 hours. You couldn't have done that a year ago.

Like 12 months ago you were still like build a workflow and then at this point in that workflow for Make.com or any of them, it would ask ChatGPT a question with this info and then you'd have the ChatGPT answer, all right, go to the next step. Like it was very clunky. It was literally building like, what was that? Software Scratch they used to play in like high school where it's like teaching you how to use, you know, make workflows and stuff. It's like it was that 12 months ago. Now it's, it's on a different level.

So yeah, let's finish with the ego part. Bigger picture, does AI make average operators feel like experts? And is that dangerous? Because what I see is a lot of people who argument with someone and you know, they're Googling it, you know, they're not experts, whether it's politics, whether it's what's happening in Ukraine and Russia, wherever, like they just, you know, everybody has an opinion. But now with ChatGPT GPT, you can get an opinion and share it as your opinion, thinking so much about yourself.

I mean, what's your actual question? Like, are people getting dumber with it?

No. Does AI make average operators feel like experts, and is it dangerous?

Yeah, I would say like anytime you give an idiot a new tool, they feel smarter with it, right? But like, I think Adam can contest to this and tell if I'm wrong, but when I started actually getting into a lot of the AI stuff, which was like 3 months ago, I've been using it for a long time, but like 3 months ago I really went all in. I talked to Adam, I was like, all right, I'm gonna commit to this. I've seriously consumed so much time. Now I feel like, A, I don't have enough time in the day. Like, beyond every single second I'm thinking, I'm like, I need to get to my desk and start working on this. But it's made me feel dumber because there's so many things that starts building and I'm like, I would have never thought to do this. I would never thought. And I'm like, oh my gosh, I would need 50 people on a team to get all of this compressed. That would take me 6 months. And now it did 90% of that without me, you know? And so I think actually anyone that's really into AI probably feels dumber now and is like taking a step back and being like, oh my gosh. Adam, have you had that same feeling?

Yeah.

And the key things that you said there was like 90%. AI is so good at getting you 90% of the way to a 100-hour project in 2 hours. Going to that last 10% is really, really tough because you're saying ego, right? People start thinking that they are able to solve problems they can't solve. And that is the ego is like, is it dangerous? I think it's more dangerous for them than it is for the people unless they're selling something, right? Because a lot of people can look like they're able to do something and then they will sell it. So you see it. Every roofing company is a tech founder now. And they're trying to like sell CRMs that you look at their website and like the reporting that they're showing is an AI-generated image. It's not actually in the software. Like there's like the blurry words in the corner and, and stuff like that. And like, these are people that are getting on podcasts. These are people who are speaking on stages, people with a personal brand. It is giving people an ego because they're not saying, I know this. They're saying, I have access to the information about this. But there's a counterargument in that I think that the general level of people to be able to satisfy a piece of information they don't know so quickly has a general uplifting effect. You see this a lot in the international view on AI. So the biggest proponents of AI are disadvantaged countries, South America, Eastern Asian groups, because they're now suddenly capable of serving these other markets. Whereas you see a general discontent with AI from America, like North America and Europe. It's a complicated issue. It really, really is. And I think that that's where that ego needs to be checked. And I think the invisible hand of capitalism will check it. And I think the subsidization of AI will do it because right now, if you pay $200 a month for Claude Max or ChatGPT Unlimited or whatever, you're really getting around $5,000 worth of AI. And that's coming to an end. When these IPOs happen, you're not going to see these crazy subsidized AI things. And now people are going to be left like not having learned anything in this incredible opportunity period. So the time is like really right now, and you want to understand the tool, not fake it that you're smart.

That's the best point you've made this whole time. Sorry, not cut you off, but that's something you can relate to, is like Meta ads at one point were so dirt cheap compared to what they are now. Before that, it was Google Ads, right? Before that, it was, I don't know, TV, late-night TV stuff. So it's like, I do agree with you, Adam, that AI as a technology, as a tool, is funded right now and so cheap compared to what it'll be, it probably will correct itself and it'll price most people out because once everyone realizes how valuable it is and it really is full scale like it is.

I think it's going to be cheaper. Look what happened to Google. Look what happened to Facebook. Like Facebook is free. Instagram is free. YouTube is free.

Yeah, but those were free to get an audience because the audience isn't their product, right? Like you using Facebook makes Facebook nothing. It's selling you advertisement.

I think their model going to be, and they're already talking about it. I think ChatGPT will be just like Google. You just see ads.

The thing is like you won't be getting 5.5 Pro and Claude Mythos 5 in your free with ads plan. I said this a long time ago when ChatGPT first came out with a $200 plan. I said, if you're not paying for the $20 plan and learning how to use this tool at $20, then you won't see the value in paying $200. And if you don't know how to use it when it's at $200, you won't see value when it's $2,000. And when it costs $20,000, if you didn't learn how to make money with it at $2,000, then you are going to see the have and have-nots of this technology. And it's going to create a wealth concentration that is very scary. And that's not like my interpretation. That's like the people who are working at these AI companies who are in like charge of safety or security. And they're like quitting their jobs and moving to Ireland to go fishing because they're like, I'm just going to enjoy my last couple of years. I'm not trying to be one of the AI doomsday people. I'm a very big proponent on AI. But like what I am saying is that the whole permanent underclass thing is a real fear. And it comes from the fact that like, just like in the high inflation market that we've had since 2008, you should invest in assets, right? And everybody said that for a long time. But how many people do? They don't know.

Let's go down the doomsday rabbit hole. I enjoy this. At one point, everyone was farmers. And then tractors came along and then everyone was office workers and then webcams came. There's been these progressions. The worry with where we're at right now is we don't know what's left for humans after AI is at full scale, right? It's going to get very difficult for there to be anything left for humans. But we also thought with that, with farming, that our creative is really what stands out. The thought process behind it, the coordinating behind it, the final 10% you're talking about. And so my idea is if you're not in that final 10% of decision maker, which is business owners, executives, or like your job is just proof checking things, you're at risk and you're going to be a farmer 20 years after the farm, after the equipment comes out. But now it's not 20 years for it to ever— I think everything to catch up. You're saying it's like within a couple of years it'll catch up.

What you're kind of seeing is the issue is like it's the ability for a person to acquire wealth, right? So caveman days, you get a stick and you're a better hunter, you can feed a tribe, then you're a farmer. But then the industrialists come out and now one man can own enough wealth creation to feed a city or a country. And then the information technologists come out and you have the wealth creation that you can own multi-generations of nation-level wealth, right? And then what happens when everything is paying a tokenized economy to 4 companies? When those jobs that it replaces, there's not an immediate replacement of those. It's the whole Luddite fallacy where they said like when the looms came out, every, the Luddites would smash the looms because it was taking their jobs. And the whole thing is like, well, they were wrong because it created more jobs. The problem is it took 2 generations to do it. We saw this with the industrialization of America where they de-industrialized, sent those jobs overseas, and nothing really came back to fill them in.

You're taking the stance that it's not going to take 2 generations, it's going to be so instant and abrupt we're not going to have time to adapt. That's the doomsday theory, right?

Instant and abrupt. It's going to affect knowledge workers like consultants right now. Just came out today, McKinsey is now like 40% of their pricing is now outcome-based because they can't get away with charging $1.5 million for research that they did in a few hours. And the customers are now saying, I'm not paying you that money anymore. And so consulting firms are moving towards outcome-based pricing. We'll take 5% of the business or we will take 8% of the outcome. That's the first stages of it. Subsidization of AI will make it so there's, there has and have-nots. People have not learned how to use the tool to generate an ROI. And then as you see physical AI come in, it's going to come down, like the government's going to try and solve it with universal basic income. And I think companies have a responsibility to look at this. That's part of the project that I'm working on. You know, why I'm trying to come back to the States is because I have a plan about how roofers as physical AI can prevent themselves from being completely displaced. You're already seeing concepts like this, a company called Human Archive.

right?

That is paying people to train the AI. But what happens when it comes in its place? There will be a period of dystopia before you reach a new level of like utopia where you see this great economic boom. There'll be a short period, but we have all kinds of other things, human elements that are gonna come in. You know, people are gonna bomb data centers, people are gonna block them from being built. People are gonna boycott AI. You're gonna have the humanists that don't want to use it. You're gonna, you're gonna see all kinds of disruptive events along the way. Look at the jobs report, right? We tripled expectations in the jobs report because AI right now is creating more jobs because so many companies are now looking for ways to take advantage of this technology. So there's this guy, Nate Jones, and he talks about the surface area of your AI skill is gonna create all kinds of really interesting jobs going forward, but there is going to be a part of the economy that gets left behind and we're concerned about What is the implications of that on society, on families, on fathers having a purpose? These are all really big concerns that I have, and I feel a responsibility to start educating about it. That's largely what this content that I'm doing is about.

Let's wrap it up. I have 2 questions. Let's do 2 minutes per question. Give me one. I know it's hard for you, Adam, but let's, let's try. Give me one AI prediction for roofing that people will laugh at today, but believe in 3 years?

Insurance companies are going to be done coming out to roofs. Meeting adjusters on site will be over within 3 years. I'll tell you, we could probably finish it right now. Like, there's no need. I mean, there's already claims we get on where the insurance doesn't come out because our reports are good enough. So yeah, within 3 years, all of your meeting a State Farm adjuster or a SignNow person doesn't actually make any decisions like that's going to be gone. AI is going to extinguish that. And so industry specific.

Okay. Bolt, Bolt. Adam.

Adam Sand will have a humanoid robot on a roof nailing a shingle in 3 years.

I don't even know if I have one. I think we're going to have a big bubble. I think economy will collapse after all 3, 4 companies will go public. I think it's going to be big money grab for all 3, 4. Massive money grab.

Bigger than internet, bigger than train tracks.

Believe it's going to crash. And we're pretty much watching redistribution of wealth in this country, which very sad to see. They're pushing it on us. The value is not what they're putting. I mean, we don't have a couple trillion dollars in AI value.

I don't know if I would say they're pushing on us, like as cynical as I am, as like, you know, anti-government I may be.

No, no, demand is there. Demand. Correct. Well, here's how they're pushing it on us. And it's actually genius business model. So what they are doing is it's actually been created in Russia in '90s. We have a Russian oligarch who created the system, and the guy who used to work with him came to United States. It's pretty like the system is very, very simple. So let's say you own a business and the business does not have funds and stuff, and you go borrow $100 million, and now you invested in this business, this business invest in this business, and this business invest back in yours. So now you create this full circle. The money is made, but no value is created.

So what they're doing right now is NVIDIA invests in OpenAI, and then OpenAI agrees to buy NVIDIA chips.

Exactly. So on paper, everybody made money. So they've done it in Russia in '90s. It was genius business. Like, few billionaires were made. And there was a Russian Jew, like Berezovsky is the last name he got. Suicided in England later when he went against Putin. But anyway, so he kind of invented it. So he actually, it was genius on paper. And his assistant came to United States and they're doing it like with all these tech companies. So between each other, so they're buying, they're creating wealth. And then to us, they're like, hey, we're worth $1.7 trillion. And we're like, no, you're not.

Didn't you just explain every PE company out there?

PE is a little bit different.

Very similar. No, no, no, no, no.

The model of PE is the trade value value of your stock, right? Like 2 companies, you know, traded 4x, and if you have 100 companies, it's 40x. So that's PE model. Here is I give you $50 billion, you give him $50 billion, and we go public as a $500 billion company. But nobody, like we're not Walmart. We don't have $500 billion. It's just like we exchange money between each other.

There's the circle jerk of the AI companies, right? The NVIDIA invests in OpenAI, OpenAI buys NVIDIA chips. There's a guy who put out a really good infographic on it. Then there's all these SPVs that have been created. And SPVs are a good thing, but they're a way to sell blocks of stock without diluting. And all of that is all financial mechanisms that are good today, bad tomorrow. Derivatives, the same thing. All these things are good today, bad tomorrow. But the fundamentals of the economy lie underneath. And what you're really seeing is you're seeing another massive infrastructure transformation. It happened in the 1870s in Europe. When they did the train tracks and happened again in America, localized to America in the 1900s when they did the train tracks because there isn't enough value created immediately by building the railroads. But once the railroads were built, then there was 100 years of economic prosperity created on the backs of the railroads. The same thing with the internet. The internet came out and then companies lost billions of dollars. IPOs went, popped and dropped and companies were big and small and consolidated. Because they were laying, you know, hundreds of thousands of miles of fiber that really didn't become useful until mobile internet, like mobile phones used it. So you saw that whole thing happen in the internet craze. This is another major infrastructure build where the fundamental of the technology is good, but the amount of investment required, right, to build the infrastructure for this, there will be winners and losers. These patterns are not new. When you zoom out and you say, what is AI? It's a major infrastructure boom. It's data centers, it's power, it's cooling. It's all, all of these things, right? If you're going to buy the IPO of OpenAI and Anthropic, go ahead, but don't buy it on leverage and don't buy it waiting to get your money back in 3 years. Because the same thing happened to Facebook when Facebook came out, IPO'd, went up for a bit, went down. How are they going to monetize mobile? 3 years of pain, and then huge again. Like, if you're buying OpenAI or Anthropic stock when they go, or the Tesla lot, the SpaceX one, like, don't buy that stock on leverage and don't buy it unless you're planning to hold it for a decade.

Love it. Any last words? Like, last words, not last sentences. I know.

If you want AI in your business, in your roofing business, call Adam. If you want AI in your claims process, call me. If you want someone to actually answer your phones, call Dmitry. We're here for you.

What an ending.