How to Use AI Agents to Do the Work For You⏐Holdco Bros Discuss! Ep. #116

Who likes calling people on the phone?

No one.

No one. That's universal. You can click a button and call a thousand lawyers over the next five minutes. To us is like, oh, it's called an agent. But to 99.9% of humans, that is mind blowing. I ended up with a spreadsheet of 20,000 repair strips. And I remember thinking, I know more about this industry than anyone on the planet right now. This is seriously a good idea. You could sell this as a service. You could sell the data to people. You could build your own newsletter. Never. Callagain.com. i own it.

Oh, dude. I'm done. Are you freaking kidding me, dude?

Well, Nick, do you want to kick us off or should I? You call it, Chris.

It's 2025, okay? I want you to kick us off.

All right? I got an idea. I got a business.

He just took it. He didn't even fight it. Okay, no.

There'S no time for pleasantries. Okay, so did you see that tweet by Andrew Wilkinson about how he created an AI agent to set a restaurant reservation for him?

Yes.

Okay.

Did you watch the video of him?

Yes.

On the tweet?

Yes.

Okay. Did you have the same thought I did.

Do we need to edit this out? I don't know.

We might. I was like, you're a billionaire. Like, he was in, like, this, like, cardboard cutout.

Yeah, There was, like, plywood behind him. Plywood. Blink twice if you good, bro. Is that, like, a Chipotle plywood vibe? Like, are you. Are you purposely intending for that to be a vibe? Or you.

Was it designed on purpose?

I did notice that. Anyway, so it's just, first of all, framework zooming out. You and I both know about AI agents. We both know that this tech is possible. There's a company called Air AI that has been enabling, like, voice caller agents for over a year now. And I played with it right when it came out, and, like, this is a very simple use case for AI agents.

Like, July 2022, we were playing with it.

Yeah, very simple use case, but just seeing it and just the context and the framing, I don't know, it just set off this flurry of ideas for me. Isn't that interesting? Like, I knew it was possible, but I just saw it and it's like, okay, that's not the end idea, like a restaurant reservation, etc. But how could you apply this?

It's because when we played with it, it wasn't good. Like, it was. You could see that it was gonna be something good at some Point. But the technology just wasn't there. And I just don't think we've adapted to how fast this technology is changing. Because in my mind, yeah, AIR AI would still be what it was two years ago, and I haven't played with it for two years. So of course you're not up to speed with, like, how good it potentially is right now.

So from what I saw, honestly, it was about as good as when I was playing with it, because what was the same for me was the pause. There's still an awkward pause in his video and when I played with it a year ago. And so I didn't see enough in the video to know how much, like, logic the AI had. But I know that the pause is, like, the worst part because it's very unnatural. Okay, sure. But you can get around that. So I watch the video, and then I go over to that room and I start doing the dishes. And I'm just thinking, I'm just thinking. I'm just scrubbing and thinking and scrap.

I'm just thinking. I got bubbly.

I'm scrubbing the glasses.

I got the silverware soaking in the water. But can we just talk about how much we both love washing dishes? Because it's like that mindless activity that allows your brain in the background to, like, run.

And it, like, it makes us pretend like we're halfway decent husbands for like, an hour.

1. It solves the ADHD of us where it's like, oh, I did a task and there's the outcome, like, really quick. You can see the result of what you did. Like, I. I do enjoy washing dishes.

The barbell strategy is employed when I do the dishes because I only do the dishes when I'm in a good mood or a really bad mood. If I'm. If I'm in a steady state, I ain't doing those dishes. Okay? If I'm angry at something, I'm going to be scrubbing those dishes so hard.

Oh, my gosh.

That's just me. But I do. I. I do enjoy it, honestly. So I'm doing the dishes and. And I'm thinking. And my kids are like, dad, what do you think about this? And I'm just like, huh? Like, dad, dad. And like, I'm just thinking of ideas. I'm in the zone.

Dad, I love you. I just want to tell you one thing. Really important. Yeah, whatever.

Dad. Mom left for milk and she hasn't been back in seven hours.

Dad, do you still love me? Because last night. Anyway, sorry. Keep going.

So who likes calling people on the phone.

Like, to no one.

No one. That's universal. You're not special if you don't like making phone calls. Right.

Talking to your friends is one thing.

But sure, sure, sure.

Something. Yeah, yeah. To get something done. No, nobody likes talking on the phone.

And so I started thinking, how could we productize something like this? Okay, first of all, remember what we did with the Stump Stump Grinding business? We had a virtual assistant call every Stump Grinding company in Houston and asked a few questions. He surveyed them. How much would it cost? Would you ever outsource it to us? Yada, yada, yada. It was really interesting. We did a couple of podcasts in it. It was like July. If you want to go back and listen, we could do something like that. We could do like a business idea validation agent that calls up businesses and asks, you know, does research, what is your pricing and all that. That was like my first level of ideation. What do you want to say?

No, well, I'm smiling because you and I never share notes beforehand, and we didn't this time either. Yeah, this perfectly integrates with, like, one of my ideas. So it's not the same, but it does perfectly integrate. I'm not going to mention it yet, but this is similar to the idea that we had a couple of months.

Ago about the car negotiations.

Yeah, car negotiations, where we'd have a VA call. It's a va. It wasn't AI, but it was a VA calling around to, like, negotiate the best prices on people's behalf. So I like where you're going. Keep going.

So that was the first line of thinking. It's like, okay, well, our audience, they're business owners. They're aspiring business owners. You know, we, you know, if someone wanted to launch a home service business, let's say a plumbing business in Dallas, they could call around and, you know, what percentage of people are answering their phone? Because it doesn't even matter if there's a pause if no one's answering. You could just mark that on a spreadsheet. How much is the cost for a toilet replacement from 20 different plumbing companies? Let's compare side by side. And then it's like, okay. But that's like. It's not even that. It's too niche. Cause I'm, you know, I'm cool selling stuff to aspiring business owners, but people just don't do this. We'd have to convince aspiring business owners to do this type of market research. You know what I'm saying? But, like, this is valuable to do, and we can help you do it. That's a lot of friction. So I quickly moved out on from that. Okay, okay. And then I start thinking like mass market. Okay. So what if you're a 24 year old woman that's engaged and you're planning a wedding? You're going to be making a lot of phone calls, venues, catering, photography, flowers.

The venue.

I knew you were going to say that. Like, the more people you call, theoretically the more money you save because you're getting more quotes, more data points. Right. But most people aren't going to call that many. They're going to get a quote that seems okay and then they're going to stop. Okay, Right. So then I start thinking mass market. And it could be like wedding planners or for people planning a wedding, it could be a service for them to call. And this is what the agent would do. It would say, like, you get ahead of the pause. And it's very important here. It's like, hey, I'm working on behalf of Mallory Smith. First name, last name in city name. Very important. Right. I'm not, you know, some random robot calling you, trying to sell you something. She has a wedding planned on March 31st. I don't even think there is a March 31st. March 30th.

No, there's a March 31st.

Okay.

Dude doesn't even know his month.

The day. I don't know my month. Hold on, let me count my knuckles. Let me count my knuckles.

Oh, okay.

Yeah.

Oh, that's a 31.

Yeah, it's a bit. It's a bit. Now, Mohammed, edit this out. But actually don't because it's funny. I want people to hear it. Don't edit that part out either because I want them to hear me telling you to edit.

Kelly was just telling me we need to have a safe word for what really needs to be edited out.

I know. Anyway, so it's like, okay, so the agent calls and it's like, hey, I'm calling on behalf of Mallory Smith. She lives here in Plano, Texas. She has a wedding plan for March 31st. She. She's trying to get as many. And it's just like the first, the intro is just like a pre. It's just a very long preamble because there's no pause.

You're telling them like, basically, don't hang up on me, I'm an AI agent.

Well, you're not even there yet. But we're getting there. Yes, because they, they probably don't know in that first preamble because there hasn't Been an opportunity for a pause, right?

Sure.

So it's like, hey, I'm helping her with this wedding. She's trying to get as many quotes as possible. I like a quote for X, Y and Z. Something very specific. Something that's like their language, right? Like three 50 heads. We're budgeting for 15 to $20 per head. We want two appetizers, a side and a meat. Right. And then it's like, also, at the end of it now, like, you've got them, they're hooked. At the end of it, it's like, also, I know this is weird, but I am an AI agent. You're going to see a pause in this conversation, but ask me anything you wish. I'm going to report all of it back to Mallory and if your quote is competitive, she's going to give you a call herself.

I knew these AI agents were stupid. I'm gonna see a pause.

Okay.

I don't know who this is. I don't see problems. I hear pauses.

Arnold Schwarzenegger's calling. It's here.

I'm calling on some wedding questions. Do I get to the chopper first or second?

Anyway, so you finish it. You finish it with like, I am an AI agent. I know this is weird, but like, please ask me any questions you wish. I am intelligent and I'm going to report all this back to Mallory and she's going to call you back herself if your quote is competitive. And then you're like, okay, wow, this is weird. Okay, well, our pricing is. Okay, well, does that pricing include silverware? And it's going to ask all these relevant follow up questions. And it's literally checking boxes in a spreadsheet. And then they hang up. And then Mallory just gets an email summary at the end of the day. Oh, wow, this one's competitive. Yeah. I'm going to give them a call and it's like the offer is the no brainer offer. Like, we will save you $5,000 on your wedding. Or like money, full money back guarantee if you don't feel that we saved you any money.

I like it being hyper niche because you can have an AI model that you can train.

So even is there like go off the rails.

Right, right, right, right. But her name's Mallory in this instance.

Sure, sure.

Okay. So it's like Mallory finds this and from the beginning it's not just making calls. It's like, all right, Mallory, have you thought about this? It's an iterative and interactive process with the AI agent to figure out what exactly she's looking for. And then based on that set of parameters, they can go out and search for the venues in their area because they're all now able to go out and crawl the open web. Tell Mallory, here are the top five that I recommend you call. What do you think? Mallory says yes or no and then it goes out and executes on that. But because it's, let's, it's, let's say it's just trained on weddings, it knows what questions to ask. You can, you can train it to ask specific date. Like you can enter in your own proprietary information if you're a wedding venue or whatever it is. I'm out of my depth with the wedding stuff.

And it's going to ask questions that Mallory wouldn't even think to ask because guess what, she's probably never been married before, you know.

Yeah. Because you're either going to pay $10,000 for a wedding planner or for most people, you're going to do it on the cheap and you're just going to plan the wedding yourself. This is the AI Wedding planner, the cloud wedding planner. Right. Like, and they help you plan everything related to the wedding and they know what questions to ask. They don't know a general price that it should be right. Like, how many times are you pricing something out? And you're like, I don't know, is that good? Is that a good price?

And yeah, you don't call the priceline.

Yeah, I can go call the web and say generally speaking, this is how much this should cost. And then the more people that you have using that product, it becomes a network effect because then you're actually getting the data that people spent. So it can get proprietary information to level set and say this is what market is for these services anyways. I, I like it, I like it a lot for niching down honestly, it's.

Like it could be as simple as a chatbot and you can integrate it with imessage with just a text chatbot. It's literally someone you have saved in your phone and you treat it like, hey, how much should catering costs? Like in, in Plano, Texas we one guy quoted 30 bucks a person, one guy quoted seven. Like what is average? And it will have all this expertise to answer you. And then you can say, okay, go call 30 caterers and I'm targeting 13 bucks ahead. Okay, I'll be back in two hours.

Okay, you glossed over this, you can integrate it into your imessages.

Yeah. With text message.

I go right now and I just open the app every single Time with Chat GPT and I'm like hey, blah blah blah blah. But if it was just integrated into my text messages, is it an imessage or is it a. Is it blue or green?

I think it's going to be green.

Nick, I will remember you.

We, the 38% of Android users are logging off right now. They just, they just put their phone down and unsubscribed.

That's. I didn't know that, I didn't know you could. How do you integrate it with your.

Got to find workarounds like I know for, for the average consumer, I mean you could use Lindy or you could use Zapier or through Twilio. Through Twilio, yeah.

That's a tweet. Anyways, sorry. I like that idea put as a text message and then you're just literally questioning it. And it's, it's not like Google in the sense that it's I ask a question, it gives a specific answer. Many times when you're using these models, ChatGPT, Claude Gemini, they'll give you an answer and then they'll ask you another question.

Right.

Like oh, did that answer or have you thought of these things or would you like me to plan X, Y and Z? Right. Like it does. It's an interactive conversation as opposed to just a one way asking answer, ask answer. It's like ask, answer, ask back.

Yeah, it's like it's Chat GPT that can go out and do things for you which to us is like oh, it's called an agent. But to 99.9% of humans that is mind blowing to them because it's like the person, the Mallory that we're selling this to. And I'm picturing two pricing plans. One to the woman who's planning her own wedding and then one to the wedding planner that's going to be a much more high volume user. They would need to pay more. Right?

Yeah, that's a good one.

All of these people like 20 to 40 year old woman is used to using Chat GPT most likely. Right. But if you were to say hey, I want you to ask it like what an average catering order costs. Okay, I could do that in ChatGPT and then I want you to ask it to go call 30 caterers in your area and report back on prices. Oh wait, what, how can it call? And then it's like well do you want me to call you right now and we'll do a sample call? Yeah, sure. They'll pick up her phone and she can play the role of a caterer and she can just ask questions and just see how good it is.

Let me ask you a question. What percentage of the time do you use Google and what percentage of the time do you use an AI model?

ChatGPT Claude I almost never use Google anymore unless I'm in Google Maps or there are some exceptions.

That's what I think. Searching for physical businesses, I think.

Yep.

And products like that I still use Google for. And I think I would bet even even though our audience is probably tech forward, 50% of them still don't use chat, GPT or an LLM in the way that they should be using it. And I want to get like this is a perfect example I'm going to give right now. We bought our son's Apple watches for Christmas. They had the little gizmos that were set where they could only call like a couple of people and they were fine for a period of time but the GPS on them didn't work really well and it was gated so we couldn't add that many contacts to it. There was a certain number of contacts that it could add. And so we were like, you know what, let's get the Apple watches for them. Cool. So we get them Apple watches for Christmas but we still don't want them calling everybody. We don't want them getting out to the, to the open web. And also we don't want them to have access to the game center. We don't want them to be able to receive calls, but we want them to be able to make anyways. There's like a whole set of things that we wanted to set and order to gate their experience. If I went to Google and I said how do I set permission parameters on the Apple watch for new watches? I'm going to get a list of articles that were pre written with somebody else's idea in mind of what they thought I might be asking. So it's not going to be exact. There may be one article that's like, hey, how to limit phone numbers.

Cool.

There may be another article that's like how to make sure the game center is not active. Or there may be another article that says how to make sure the web isn't available. But Siri still is. Right, whatever.

It's a lot of clicks. Right.

It's a lot of clicks and I've got to read and I've got to then interpret, internalize and then I got to do it myself. I went to ChatGPT and I wrote a long prompt and I Said, here are the things that, exactly that I want. Please give me a step by step process to do this. And guess what? It was perfect.

Yeah.

I didn't even have to ask it a second time. And Jess was sitting there with me and we went through it and it was like, okay, go to your phone, open this setting. Okay, click on this. This should be, this should be here. Go to the watch. Very, very easy. And I think what most people don't realize is that change is so enormous. From hey, here's an article that might help you solve this to this is the answer. And then what we're talking about on top of that is this is the agent that will go out, execute this function and bring it back to you with an answer. It's two step functions away from even Google at this point. Even the people who are listening to this I think probably aren't using ChatGPT enough.

Yeah, I mean, you can take it even a step further. Like I, I was making sourdough bread for the first time and it's a little complicated and I pulled out my starter and it had like deflated and I made the sourdough and I'm like, okay, the consistency is kind of weird and we're just having this conversation. And then I took a picture of the sourdough and I'm like, what does this mean? What does this mean? And it's like, oh. And what it told me is exactly what it was. It was like, oh, it's because you didn't leave it out long enough. You need to do this. You need to feed it with 50% flour, 50 water. And it's like, how many chefs or how many people cooking in their home kitchens are even thinking to do this?

They're not.

There's probably 20, 30 apps in the App Store right now, like AI enabled cooking apps like this, that offer this. I bet you there's room for like 200 more. Because it's so early.

It's so early.

If you're listening to this and you have an idea like this, like, oh, like a, a cooking assistant app that's $5 a month that you can talk to and take pictures of. And then you go, you're going to go to the Apple App Store. I've never done this. I just know you're going to open it and you're going to see dozens of these and you're going to be disheartened and you're going to move on. Don't be like, press forward, cursor. AI is like the website for specifically building iPhone apps using AI and it's like plug and play. So there's just room for so much more.

Have you played with the chat GPT? I'm sure Gemini does this as well where you can actually you just talk to it, ask it questions.

Like verbally. Yeah, when I'm driving sometimes I will.

But I'm not trying to over exaggerate. It's. I feel like I'm living in the future because. And I have to do this. I know this is going to sound so stupid. I have to tell myself I have access to whatever, to whatever I'm curious about. Let's say that I want to learn about Bitcoin. It's like all right, I have access to Satoshi, I'm going to ask him questions and so I'll ask a first question and chat CPT will give me the response but then I, I can ask follow up questions. So if I'm on a walk or if I'm driving I can have a conversation and it's like teaching me as we're talking as if I have, you know, Satoshi or whoever the PhD of the topic that I'm most interested in and it's giving me real information. It's like a, it's like a course in my car which is, it's just nuts to me that I have access to that right now. Like there's so much ability to learn.

It's crazy. Part of opportunity. No.

So I love that idea.

So I bought a domain name for it. I bought a domain name for it. So I think you could do this with the generalized name I'm about to tell you and just focus on weddings and then go broad from there. Or I think this could be a standalone wedding product like we talked about. But the domain name I bought. Are you ready?

Yeah.

Have you heard of do not pay.com?

No.

So I mean Do Not Pay is like it's a nine figure company and they started by just like it was an automated service where you could dispute your parking ticket in San Francisco. That's how niche they started.

That's amazing.

Yeah. Now it and they've been funded by like Sequoia and all the big names. Do Not Pay dot com. I'm trying to see like what their purporting themselves to be your AI consumer champion. Do Not Pay uses AI to help you fight big corporations, protect your privacy, find hidden money and beat bureaucracy. This is a big company, started out very niche. So I was thinking along those lines. What could I name this? That's like, it just stands out. It says exactly what it does. A few syllables. You could start at niche with weddings and then go as broad as you want, like they did with parking tickets. Never call again. Dot com.

Oh, my gosh.

That's it. I own it.

Oh, dude, I'm done. I'm done. Are you freaking kidding me, dude?

He's. He's gone. Are you freaking kidding me, dude? It's a good one. It's a good one.

Do not call again.

Never call again. Never call again.

Never call. Oh, gosh dang it, Chris. It's so good.

Good, right?

It is so good.

So it could also be like a monthly membership, like, just to keep in your back pocket. 20 bucks a month. It's like, oh, our AC just broke. I want 10 quotes. Click, I got 10 quotes.

But that name, though.

It's a good name.

But that name, though, gosh dang it. That's a good name. Never call again dot com. Holy crap.

So anyone listening to this, you're welcome to go copy me, but you can't. You can't take the name so well.

I think this is going to be an all AI episode.

Oh, shoot. Okay, we're doing this. I mean, how could it not be? It's so transformative.

It is. It's very trans.

All right, if you are watching this clip right now, you are watching this on YouTube. And if you are a fan of podcasts, please check out my podcast while you're at it, because sometimes I have podcasts that aren't on YouTube, and sometimes I have YouTubes that aren't on podcasts. So check out the Kerner office on any podcast platform. Or you can go to tkopod.com Also, my newsletter will throw that in there. There are no ads. It's weekly. It's different content than my podcast and YouTube newsletter.chrisjkerner.com you'll see the link, and we'll see you there.

So is it my turn? Can I. Can I tell an idea?

Yes, please.

All right. So I was talking to my. My son about.

Or Royce.

I was talking to Royce about AI and why it's important. And we're. You know, we've been playing around with it, and I was like, royce. You know what's interesting is that knowing the right question to ask is as important as whatever answer you think you want to get.

He's like, really, Nick? Real quick, I had this exact conversation in the last week with my kids.

Really?

The exact conversation. You've got to know what questions to ask. It could open up the whole world to you.

Yeah. So we're sitting there, we're talking, and, you know, I'm telling him, royce, you're curious. I'm curious. I can't tell you how many times I've asked a question, and everybody else who's sitting around me has been like, come on, let's. Stupid. But when we start drilling into it, it's. It's like, oh, nobody really knew the answer to that question.

Right?

Like, nobody. And nobody had thought to ask it because, oh, well, everybody knows the answer, or it just seems so obvious. And as we were talking through, I was like, did you know that there are jobs out there that people are paid a lot of money for where they're literally paid to figure out the right question to ask for these LLMs to return the right answer? He's like, really? Like, yeah. They're called a prompt engineer. It's like, well, what do they do? It's like, well, these. These LLMs are very specific, and so you have to be very specific with the parameters that you give them. What does the outcome look like? What information can they use and digest? Can it be inferred? Does it have to be cited? Where are they pulling the information from all of those things? And so in order to get the right answer, you have to ask the right question. There. These prompt engineers, he's like, wow, that's really interesting. And so he starts researching that, and I started thinking about this. Do you remember, I don't know if they still do it, but do you remember in the Apple stores where they'd have, like, Saturday morning classes and every grandma would go, and they'd bring their iPad? They still do, like, they still do it. Okay. This is how you access photos. It's like, are you my grandson? You know, and they've got people telling them what to do. I think there's a world where an entrepreneur, enterprising entrepreneur, could either do a physical in person or virtual class for boomers on how to use ChatGPT. Because right now, I mean, even teenage kids that I've shown how to do it, they don't know how to use it. They use it for novelty. They're like, oh, can you spell penis? You know what I mean? Like, it's just like, they think it's something funny.

That is pretty funny, though. I mean, come on.

It is. It is pretty funny. But they don't. But they don't know. Like, okay, what else would I use this for? Like, what's. I can ask the first question, but what's the Next question. How do I get the most out of it? And so if I was in the space, if I was in college in particular, and I wanted a side hustle, I would launch a business around educating adults. It doesn't even have to be boomers, just adults on how to ask and how to get the Most out of ChatGPT. What do you think?

Absolutely. Like, I'm thinking of other variations of this. This is just a thought I had while you were talking. OpenAI is not about to spend millions of dollars pushing paid targeted ads towards boomers that don't know how to use ChatGPT. They've got too much on their plate. Or like. Or like a boomer that wants to cook better with chat GPT. There's, like, no one at OpenAI that's not on their radar. Right? That's why you want to build a rapper that can help boomers cook better with ChatGPT. And so I'm thinking, like, a twist on your idea is to just make a ton of rappers that make the prompting easier, that basically prompt the old people to ask better questions. So, like, you download an app for 9.99 and that's pushed to the boomer through paid ads. The boomer downloads it, and then the app starts talking to them and, like, basically pushes them along. What do you want to know? And it does the same thing as what you're saying, but through an app and not through a class.

So I like the wrapper. I really do like the wrapper idea. But the reason I like doing this as a class is, is because of the reasons. You know, when you post stuff to Twitter and people are like, it's not that easy, bro. Actually, what about insurance? He didn't think about the liability that's going to come, right? And. And you. Why don't you do it?

If it's such a good idea, why don't you do it, bro?

Like, every time you post, if I had to start over again and make $100,000 this year, this is exactly how I would do it. Step by step. And then I text you, I'm like, but could you do that, please? Could you do that, please? Anyways, I think that someone with would hear this. Oh, put a wrapper on it and be like, oh, yeah, I'm gonna put a wrapper. And they start doing a little bit of research, and there's friction, and they're like, it's too hard. I don't. I see what you're saying.

For the person listening that wants to start this, that might Be kind of a bridge too far.

Yeah, yeah. But I think if they started this within three months, they're gonna, they're gonna be so good at ChatGPT. They're gonna be like, oh, I could build a wrapper. I could just have Chapter even tell me how to build the wrapper. I have all these tools and resources to build the wrapper. I think that's the iterative next step. Where I would start with this is. And you might. This might be too controversial. I would go to my local church that I'm a member of and just say, for free. I'm not, I'm not asking for money.

Can you loan me money for this investment scheme? Is that what you're. Okay, never mind.

No, no, I'd say, God told me I'm supposed to start this business. That's what I'd say. I'd go, I'd go and I'd say, hey, look, we do activities. One of the activities at nights I'd like to do is just show any member of the ward or any member of the congregation, any member of whatever parish how to use ChatGPT. Can we put on the books? And I would just test it and I would just see what the response is. Like, did it work? How was it received? Odds are somebody in the, in the congregation is going to be like, could you come and help me do this a little bit more this week? Right. And you'll get even more experience doing that. And then at some point that, that might blossom out into other business. But I would, that's where I would test it first before I went full bore into it. And then from there, I would use the same distribution channels that we've been talking about since the beginning of time. Email, Twitter, social media, etc. But if I didn't have those things, again, potentially controversial. I would go to assisted and independent living facilities, mainly independent living facilities, because assisted living facilities, you get people there who just are like, really aren't all there. Independent. They're still all there. They just need a little bit extra help. And I would just offer, offer a free class. And then I guarantee at the independent living facility, after the class, you'll get 10 people coming up to you saying, I'll pay you 50 bucks for a half of an hour if you teach me how to do X, Y and Z. That's where I would start this business. If I didn't have a social platform.

Do you think it's. That's. The revenue model is charging for your time, like, per hour? Is it like A subscription back to.

What we were talking about. I think to get into it for like, just a person listening who's like, oh, I want to try this out. Yeah, it'd be for per hour. I think eventually there's a million things you could do, right? You could create the wrapper, you could have a subscription, you could. I mean, we talked about this when we were talking about recording stories in these facilities. You could create a video of yourself interacting with them and, you know, set these up around the city for people to have, like, interactive conversations. So, yes, like, it is you, but it's not really you, you know, But I think the first step would be like, all right, prove that you can generate some income, learn the product and the education piece of it, and then scale.

I like a lot. I think a lot of. I think it would lead to, like, more higher ticket things, just naturally. Right. Totally based on the type of demographic that you're. That you're teaching to.

And the reason I like it for somebody younger or somebody in college is because I think it's going to take six months to sort of present itself on what the opportunity is. I don't think it's like heir apparent right now. Oh, this is exactly the business model that it's going to become. And if it's somebody like you or I and we leave our jobs, it's like, I don't necessarily have six months, you know, if somebody listening to this, they don't necessarily have six months to just kind of mess around and figure out what it's going to be. And then another six months until it's monetizing to a point where they can have income to pay their mortgage and pay bills and all that other stuff. So I think somebody young could do this.

Yeah, I like it. It's a good one.

Oh, Chris, I have one more idea on this. I just talked to Craig Fuller. He's the. He's the founder of Freight Waves. And if you don't know what Freight Waves is, it's a really cool business. There's a million podcasts out there you can go listen to. He's on my first million and others. But essentially, do you know what a Bloomberg terminal is? Okay, so Michael Bloomberg, one of the richest men in the world, made his money by selling information faster than anybody else. He set up these terminals around the world. They were called Bloomberg terminals. They're this big hardware device, and they had proprietary information that people used to make investment decisions. And that's where he started. And now he has a media company that you have Bloomberg Markets, you have Bloomberg the publication, all that other stuff.

Super expensive to use, right? The terminals.

Yeah, very, very expensive. Very expensive. Think of Craig as like the Bloomberg for logistics and shipping. So he has a media brand which is freight waves, where they put out information industry publications, etc. And then they've got a Bloomberg like terminal in. Sonar is the name of the company. But the idea was that they're going out and they're aggregating all of this information and they're giving it to people who wouldn't have the ability to go out and aggregate it otherwise. All right, it's a hundred million dollar plus company, totally bootstrapped, owns it 100% anyways. It's a really, really cool business. As I was talking to him though, I started thinking what could we, or what could somebody build around this thesis which is unique or proprietary, that's the better word, proprietary information. And what could some of those be? One might be searcher sentiment. So people who want to buy a business that are out there searching actively right now, ETA is a big deal in the, in the SMB world.

Entrepreneurship through acquisition.

Thank you entrepreneurs. ETA entrepreneurship through acquisition. What's ETA searcher sentiment like right now? How would you, how would you get a survey of all the searchers? How would you get all that information? Small business, pizza restaurants, I don't know, how would you get a sentiment of what those business owners are feeling right now? And the reason I started bringing it up is because Craig was, Craig like predicted the last four recessions before anybody else was predicting them because he had the logistics information, he knew what was shipping from China, he knew what Walmart was ordering, he knew the inventory was going to be on the shelves because they had this proprietary order flow that they were tracking. So it got me thinking about kind of what could you build around that in small businesses? And back to the AI agents. I think if you had AI agents going out and just calling, making the call and aggregating the information on your ideas suck.

But this one's a good idea. This idea right here, bro. This is the 1 billion.

It's a billion dollar idea, bro.

Multiple 11 figure idea.

I can't even count as high as.

It'S worth, bro, I'm telling you.

But if you had, the hard part about that business is, is you're an information business so you have to aggregate and collect the information. Craig was able to get the information because he's plugged into all these systems and he's getting like real time electronic communications for what they're ordering.

Right.

It's hard to do that with sentiment. So AI agents going out, collecting sentiment, aggregating that, and then either selling it as proprietary access, or you, as the business, are making investment decisions based off of what, you know ahead of time. I mean, there's kind of a hundred use cases, but that's the idea. What do you think?

I can't listen anymore because I love this idea so much, and I'm just thinking of ways to apply it, so. All right, couple schools of thought. Okay, let me back up. So I was thinking of when I started LCD Cycle, right? And for those that don't know LCD cycle, I sold iPhone parts to iPhone repair stores. When I started that business, I started it by scraping every single iPhone repair store in the country. And it was very tedious. And I ended up with a spreadsheet of 20,000 repair stores. And I remember thinking very distinctly at the time, I know more about this industry than anyone on the planet right now. Who else is doing this? Who else is scraping every iPhone repair store and then analyzing the results? And then I called them. Or when I say I, I mean three virtual assistants called them and started a $10 million business all built on that one spreadsheet. Okay, you talk about one chart businesses. I'm talking about one spreadsheet businesses.

I'm gonna give you one more example that you're not gonna say. Chris, a couple of years ago, partnered with John McAfee. Yeah, the crazy John McAfee. They launched a token. It was called no BS Crypto. The whole thesis was Chris was crawling Reddit, Instagram, face, Twitter, whatever it was, right? And he was getting sentiment about coins from places that nobody had looked. And it was true. It was a very accurate predictor of which coins would pop. And he got a meeting with John McAfee. You can go watch this video. Chris has recorded it a couple of times. He even got a meeting with John McAfee. John was like, nobody could predict this. Okay, show me your model. And within 20 minutes, John was like, chris, I love you. Will you marry me? And they went in. They went into business together.

But I meant, show me your female model, like Crawford.

But, yeah, I mean, you did that with LCD cycle. You did that with no BS Crypto. It's about five one spreadsheet business. It's a one spreadsheet business.

Okay, so, okay, so my mind went there, then I went a level deeper, and it's like, okay, okay. How you want to find the sweet spot between a business owner that will answer the phone Right. With few gatekeepers and a high enough ticket size. So let's say convenient convenience stores. Right. Let's say we want to analyze the sentiment of convenience stores because there's 250, 000 convenience stores in the U.S. it's like the biggest retail industry there is, second to restaurants. Okay.

Convenience being like 7 11amp. Yeah.

Like just a freaking gas station. Yeah. Like independent convenience stores.

Yeah, yeah, yeah.

A lot of the independent ones. You're going to call and you're going to get the owner. Right. Or like a nail salon. You're going to call and you're going to get the owner. There's no friction to get to the owner. Okay, I'm listening. I just, I knew you were staring.

At me the second it was like 30 millisecond. No, I was here.

I was here. So that's great because you can get to the owner and you could, you could build a big business off this. You could say, hey, I'm going to ask you five questions about the sentiment of the industry, whatever it is, let's say convenience stores. What, what are your, like, what is your sales velocity recently? Like, they got to be broad enough questions to where they don't feel like protective of it, but they also have to be specific enough to where it's good data. And then you're like, hey, and in return, I'm going to send you an email with a report of what everyone else is saying about this. It's such a perfect offer because I need 30 seconds of your time. But I going to spill the beans on what everyone else is answering me. Like, why wouldn't they want that? Of course they would. So that's great because there's 250,000 of them. You can get to them easily, but it's not very high ticket. Now in a sense it is because they're business owners, but what they're selling is not high ticket. So then I, I, I go to the other end of the spectrum. Like dentists. Okay. Tens of thousands of dentists out there, but you're going to have to go through a receptionist. I am. We all are in a sense. You got to go through the receptionist. It's really hard to get a dentist on the phone, but it's very high ticket. You can sell high ticket stuff to dentists. So then I thought lawyers. I've done this before. I've scraped a list of lawyers. You can get their cell phone numbers quite easily. There's not as much friction to reach them, but it's very high ticket. And you can, there's a hundred thousand lawyers, probably much, much more. And you can just call them and ask them a list of questions, then say, hey, where can I email your answers to? Not your answers, but everyone else's answers aggregated. And then you opt them into your list. Now you have a newsletter that like your CPMs cost per mill, aka how much you can charge for advertising on this newsletter are insanely high because it's all lawyers reading it. In order to get the newsletter, they have to answer the questions. In order to answer the questions, they have to get the newsletter. And so you're aggregating the data and building a newsletter at the same time. And then you can still like, it's not all you have as a newsletter. The data itself is incredibly valuable outside of the newsletter. And now you can do this with agents. You could click a button and call a thousand lawyers over the next five minutes.

Everything that's done right now when they're going out and polling people is done off of statistics modeling. The more people that you poll, the more answers that you get, the better your model is. However, there is a selection bias, right? So it's like the, it may be that the people who are willing to answer skew the data, et cetera, whatever, but the more data that you can get, the more accurate of a picture it becomes. And when you're looking at these smaller industries that are niche, that don't have as many operators in them, there's not an economic incentive for you to hire even somebody in the Philippines to call these people and get answers. But if you have an agent and the cost is effectively 10 cents a call, all of a sudden there's an economic incentive there and you can arbitrage the knowledge gap. And that's where I think it's like super interesting to start thinking about some of the ideas there. And again, this was from Craig and his original business and the Michael Bloomberg esque terminals of getting you information timely and applying this down. So I really like that idea.

This is a really good idea. So the downside of this idea is the cost. So I've been pricing this out recently and you're probably going to spend about 15 cents a minute for these AI agents and they round up. Some do, some don't round up. And so it doesn't seem like a lot on the surface, but that can add up over time.

What if it was like a 30 second call and you just, you knew, you knew that there was going to be friction and drop off and it was like, hey, I'M an AI. You do the pitch. I'm an AI agent. I'm gathering information on your industry, I anonymize it and then I send you the report of a thousand people who have also responded to this. Are you interested?

Yeah.

Great. What's your email? I'll send you the survey. Oh, it's john bigsby@aol.com. perfect. I'll see you soon.

Boop.

Hang up. So it's like you're not even doing the survey on the phone. Again, that would bring the cost down, but also there is friction. However, you do get their email.

Boop.

And it's a lot cheaper for you to send out that many emails as it is to like keep them on for a five minute survey as opposed to them filling out the survey online. So that could be a way around it.

That's a good idea. Yeah. The, the hard part would be like transcribing the email wrong.

There's going to be an error rate or and like there's going to be a big percentage that just doesn't click through. Like they're not going to answer the survey. You know what I mean? Like yes, there's a number of things that it could be.

You just have to, you have to ab test it. Do both two different scripts. One just ask for their email and the other one asks the questions and compare the open rates and the costs and the outputs.

But you could also tier your pricing, right, like to the end consumer. Hey, we've got, this is the cheap option. We don't do a phone minute, a five minute phone conversation. We send out the email or we get really high quality data for people who are willing to stay on the phone. And that's, you know, 10x the price. You know what I mean? So it's like okay, $100 a month or do you want $500 a month?

Yeah, you could do this with election polling too.

You could do it locally. You could, you could be like there is a city ordinance that they're going to pass and there's a vote for it by the city council. What's sentiment like? Does that make sense? Could I get people rounded up and advocating on my behalf? Like there's a, there's a million kind of applicant local applications that it could be, oh, I'm running for the school board or my son wants to raise money, whatever student project.

Are there even that many municipalities you could sell this to? Well, this is like 10 cities in America. 20.

That's a little higher. It's a small 30. 30 to 50.

Small market.

Yeah.

Hey, let's just delete this part of the episode. Muhammad, actually, don't. I'm joking. This is seriously a good idea because you could sell this as a service to people. You could sell the data to people. You could build your own newsletter with it. Like, oh.

Yep.

It's. The whole dynamic of the AI caller changes. When you're. You get ahead of it, you're just.

Like, hey, when you just own it. I'm an agent.

You own it. There's gonna be a pause. There's gonna be a pause. It's awkward. It's kind of like what we've talked about with building, like, marketing agencies, where it's like, hey, I'm a lot cheaper. I'm not 3, 500amonth because I use AI. Yeah, I use AI.

If when we were using Air AI two years ago, if we would have gotten ahead of it, hey, just wanna let you know, this is an AI agent. The call is gonna be a little weird. There's gonna be some pauses. Nobody would know what the freak you're talking about. What AI. What are you talking about? AI. Now, even if you use the word agent, they're still not going to know. But if you say AI, they're going to be like, oh, okay, this isn't a real. I think 90% of people are going to understand that there's a lag there. Whereas two years ago, there just wasn't an education.

I would probably say, like. Because they're not even going to know what agent means. I think I would say, like, it. It's like. It's as if you're talking to chat. GPT.

Yeah, that's what I mean. I would just say, like, hey, I'm AI. I'm not a real person. There's going to be a little bit of a lack of. Agreed.

Like, I'm an interactive AI, so, like, ask me anything you want, but I'm going to. You know what I'm saying? Right.

Agreed.

It's a good one.

Thank you.

Thank you. All right, what do you think? Please share it with a friend and we'll see you next time on the Kerner Office.