404: Close more Solar Sales with AI Tools for Door-to-door: How Reworked AI Elevate D2D Strategies w/ Targeted lead generation & AI-powered lead scoring | The D2D Podcast

Discover how AI can revolutionize your door-to-door sales strategies in this episode with Fred and Shyam from Reworked AI. Explore how their machine learning algorithms, combined with demographic and property data, significantly increase conversion rates and sales efficiency in roofing, solar, and other home services. Learn how Reworked AI's predictive sales intelligence identifies high-probability targets, optimizes lead generation, and enhances overall sales performance.

Fred and Shyam discuss the evolution of Reworked AI and its practical applications in door-to-door sales, from targeting the right neighborhoods to understanding customer demographics. They share real-world examples and success stories, illustrating how AI-driven insights enable smarter, data-driven decisions and streamline the sales process. Join us to discover effective methods, the benefits of integrating AI with systems, and strategies to maximize your door-to-door sales resources.

You’ll find answer to questions such as:

• How can AI improve door-to-door sales performance?
• What are the benefits of using AI in roofing and solar sales?
• How does machine learning enhance lead generation for home improvement services?
• How do AI algorithms predict high-probability leads?
• How can AI help in targeting the right neighborhoods for door-to-door sales?
• How can AI-driven insights improve customer conversion rates in home services?
• How does predictive sales intelligence work?
• What is machine learning in sales?
• How does demographic data affect solar sales?
• How to make data-driven sales decisions?

Get in touch with Fred and Shyam:

fred@reworked.ai 

shyam@reworked.ai 

admin@reworked.ai 

https://solar.reworked.ai/d2d

Thank you for listening! Don't miss out on future episodes! Subscribe to The D2D Podcast on Apple Podcasts and Spotify.

Follow us on Facebook and Instagram. You may also watch this podcast on YouTube!

You may also follow Sam Taggart on Facebook, Instagram, and TikTok for more nuggets on D2D and Sales Tips.

Transcript

00:00
don't bother contacting these people because these people are not going to install solar. That's a low target environment. Trying to predict what's going on in that owner's mind is difficult. 65% of our customers in the real estate world, they use it to scale. I had not imagined roof size to heavily correlate to the outcome. She's not telling you where the needle is. She's just saying there's no needles over here. We're 93, 94% accurate based on the feedback that Betty's getting.

00:29
Rather than knocking on 100 doors, they're going to knock on 80. Betty is using up to 50 data points when she's looking at a property. It's fundamentally an efficiency game. Machine learning can allow people to make smarter decisions.

00:47
Hey, what up D2D World? I am here with Shyam and Fred from Reworked AI on the D2D Experts podcast. Welcome guys. Thank you, JP. Thank you. Yeah, I'm glad you're here. So we're going to start by just getting into what Reworked AI is, and then we'll take the conversation on how it can affect the different industries that we're in. So who wants to explain to our listeners what Reworked AI is? So we've been talking about how best to do this, right? And if I could have 20 seconds to walk through

01:16
our meet cute, right, how we met. About five years ago, I was in land flipping, still doing that. I sent Shama a mailer because he owned a piece of property outside of Austin, Texas. He calls me and I'm giving him the spiel, you know, we can close fast cash. He said, look, full disclosure, I'm not selling this property. I just bought this property. This is a long-term hold, but I'm getting all these letters. So I explained the whole process and he said, but there's got to be some software that exists that could filter out somebody like me.

01:46
And I said, well, if there was, and actually he even mentioned that it's, you know, using machine learning. So fast forward a year, we put the company together. We're using machine learning to take data, raw data and make it actionable information. Right? So we started in the land flipping business, got into the single family, multifamily commercial properties. And just over, just under a year ago, we started building algorithms for solar. And now we're building algorithms for

02:15
for roofing. So really what it helps people do is take data, whether it be leads that they've got opt in, or they bought it from a wholesaler, and says it takes this data and uses that data as well as augmented data, and it runs it through machine learning. So if you took someone with your 30 years of experience, you would be able to identify that's a good potential place to go find a deal, to go sell them solar.

02:44
to change their review, you would be able to look at all these data points and say, yes, that's what rework.ai does. The algorithms take the collective knowledge based on history, based on actual installs and actual changes and says, this is a client that is more likely to convert and more likely to be interested in having a conversation about your product. Okay. So let's, I know we've had a conversation last week prepping for this call.

03:13
My brain still, I think AI is new enough that some of us old timers are having a hard time processing it. So let's go to the real estate piece and maybe try to explain. And by the way, when you were telling the story and you said he called, I knew immediately he wasn't calling to, he was calling for yet another reason. So in that industry, you were sending mailers based on some form of list. And you happen to send a mailer to somebody who just bought the house.

03:44
So it had been a property that you had a few months for, call it six months, right? And when I first got into land flipping, I was trying to scrub. I was trying to, you know, only out of state owners, time of ownership. But what we discovered is we all have personal bias, right? So I'm thinking that an out of state owner, somebody who lives next door to the property is not gonna sell it below market, off market, below retail. They're just not going to. But that's not the single big term is...

04:14
The biggest factor is not where they live. It's a combination of all these factors. And so when we started looking at the data and looking at my history, we realized we had to bring in the demographics. We had to try to predict what's going on in the owner's head. And so I'll ask Sham to talk more to your original question is machine learning, right? Machine learning has been around for 65 years. Right. This is not new.

04:39
This is not a large... But the machines used to fit in rooms like this. Right, absolutely. Or they were dials, right? That movie with Alan Turing, the imitation game. That's all predictive machine learning. Now it's obviously faster and easier. But Sean, you've got a great analogy that you talk about teaching a child the difference between a cat and a dog. Yeah. I mean...

05:06
So machine learning, the easiest example is one of, you show a bunch of pictures to a kid and say, hey, this is what a cat looks like, this is what a dog looks like. 10 minutes, the kid knows exactly which is which, right? So you've taught the kid which is which. Machine learning is very analogous to that, where you give it a lot of data and you train it to say, okay, here's what...

05:34
This is what a cat looks like, this is what a dog looks like, and it can then predict based on any new information that it gets, whether that's a cat or a dog. So another good analogy that I've recently begun to use is, I don't know if you've heard of Waymo. Waymo is an autonomous car company, and we've been going to conferences in Phoenix, and we've been taking a lot of self-driving cars there, and Fred always freaks out sitting in those. Much to my chagrin. And it's really, really good.

06:01
Like you see there's no driver there and the steering is turning on its own. But anyway, so the way it was going, it hits a pothole on an intersection, right? It learned right at that moment, I hit a pothole at that particular intersection. So using machine learning, it actually has learned that it did that. And also it told all the other cars in its fleet that, hey, watch out, there's a pothole in that particular intersection. There's no programmer in San Francisco going.

06:32
watch out for that, you know, part of that intersection. Right. So there's no code being written. Right. So it's, it's kind of learning on its own. There is an app in San Francisco that will tell you where people pooped on the street. Yeah. Which is everywhere. Yeah. So yeah. So that, you know, that's basically what machine learning is. And that's what we've tried to bring in, you know, to, and going back to the story that Fred was describing, right. So when, you know, when he sent that letter and I called him, I was like, Fred, would you be open to giving me all your data? Cause data is very crucial to build that model. Right. And he's like, no, let's, let's.

07:01
So to bring it to the solar world, when we decided in August, September of last year that we felt there was an opportunity in the solar install world, the easy part is building the algorithms and putting the algorithms together. That's not the most difficult part. The most difficult part is the data. So you've got the algorithms together. You've got the machine, and we call her Betty. Now you've got to feed Betty.

07:30
houses all over the country and we told her this house doesn't have solar and this house does have solar. Then she's going out there and she's pulling the demographic information. She's pulling the characteristics of the house, the characteristics of the roof, the characteristics of the shade of the roof, and she's building a scatter plot for each of those properties. And then as she's compiling these millions of lines of data, she's seeing the scatter plots that have installed solar.

07:59
and she's comparing them to each other. And then there's a grouping, a commonality of why somebody installed solar. It's not just one single data point. There's a lot of companies that talk about, well, we do some predictive understanding of what this house could be. You can't make an accurate prediction based on one or two data points. Betty is using up to 50 data points when she's looking at a property. She's talking...

08:24
She's thinking of the owner of the property as well as the structure of the property, both the economics, their history, marital status, all these things that, thank goodness, in the United States are available for a fee to make a decision as well as the structure of the property. Yeah, because when we built the solar model, on the real estate side, we were using demographic information as our core data to actually feed the models. We did the same thing with solar. And right off the bat, you know...

08:54
we got like a high 80% accuracy, which is good, but not great, right? So- It's not commercially viable, is what you're saying. And then we realized, hang on a second, there has to be a lot more data to be fed into actually build a model. And that's where we came up with the roof characteristics, right, and the billing characteristics. So all of these combined, we're seeing a lot more higher accuracy than even what we saw in real estate at this point in time of our company. So to the point of the-

09:23
The millions of data that has trained Betty, and now we're done with testing, my daughter, you're never really done with testing, now we've got customers using it, and the feedback is the high 90s. I mean, we call it 99, but it sounds dodgy, right? But it is that good to be able to say that person is your likely prospect candidate. So the data, right? And obviously, I'm probably gonna ask some questions here that you may not wanna give us the answers, but I'm thinking, okay, you may be able to tell me

09:52
Because a lot of times things that are sold door to door, there's a lot in a particular neighborhood or particular city. Bakersfield, California might be one of the most well-knocked cities out there because it's warm in the winter and it's in California, but it's in the middle of the valley and the demographics works well for selling products there. And that's why there's a lot of product there. So you can...

10:18
find data that says there's solar on this roof, this is a south-facing roof, it holds 15 kilowatt system. This roof is an ideal house, right? And they have neighbors that maybe got bought solar. But what kind of data do you get on the homeowner? You know, are you looking at like, are they searching for solar online? Or is there electric bill at a certain point? Or do they drive an electric car? Like what, like, and this might be proprietary, but I'm thinking, how do you...

10:46
great, you can tell me a home that is south-facing roof that has no shade, but that I still gotta go convince the homeowner. And that's key, right? And trying to predict what's going on in that owner's mind is difficult, right? Because they could be having a good day, they could be having a bad day. And we took a lot of our lessons learned from the real estate world, because why would somebody want to sell a property without bringing in an agent and understanding the characteristics and the value? And just...

11:15
They get a letter and they sell it, right? They get an offer and they just sell it. So it's predicting what's going on there. And yes, of course, there's some secret sauce in there, but it starts with your credit score, your credit score range, the movement of the credit score, right? Marital status, children, houses own, properties own, membership, subscriptions. That is indicative because Betty has looked at several million rows of people that have converted.

11:44
she sees, okay, these people that have bought solar, their demographic scatterplot looks like this. So then when she looks at yours, it's very far away. And this is nothing more than a predictive analysis. This doesn't tell you where the needle is, it just says it's probably not over here. And so it's the demographics of the individual. There's 26 demographic points and then another 24 that are on the structure. You've obviously seen like, you know,

12:13
I'm going to be funny here, but there's probably something that surprised you like, wow, 82% of these people who have solar are left-handed, right? They're probably not data points that you were expecting to see. Is that a safe assumption? Absolutely. So the first thing you do when you're trying to build a machine learning model is you look at correlations. Right. So you look at correlation first, and then from that, that is what feeds into the next step, which is okay, in the machine learning world, we call them as features.

12:43
what data points, essentially that's the same thing as features. Okay. What features do you use to build the model? And to be honest, I was very surprised with when we were building the solar model, roof size. I had not imagined roof size to heavily correlate to, um, you know, to the outcome, right? But, but it did. Um, so imagine larger roofs. So that's the thing like, you know, or roofs of certain size.

13:11
You know, when you build a machine learning model, you, I'm not telling the model, or the model is not telling me, both ways, roofs of size X, Y, or Z, right? All I know is there's a high correlation, and it is using that as one of the many features, right? To finally predict the outcome, but not a specific number. It would have been a specific number 20 years ago, when you didn't have machine learning, and you'd be like writing a code, say,

13:41
if owner greater than 30 years old, right? Like blah, blah, blah. Like you'd actually write those if statements, but that's not, doesn't exist in machine learning. Okay, so the data points that are giving us predictability of somebody that's most more likely to say yes to solar than no, they're just these correlations that have been added up over millions of homeowners that have bought solar. Right, and the thing, when we first put Betty together two and a half years ago to start writing the algorithms.

14:10
Obviously, I was not writing the algorithm. Shaman and the team are writing the algorithms. I was sharing my feelings of what made somebody sell a property. So then we started to look at the data. But I wasn't right. It was a personal bias of what I thought. I thought out-of-state owners, older, lousy credit scores, just to pick three. Not really. It's not necessarily the case. It's more of the credit scores that are.

14:38
on the top of the bell curve, not the really high, not the really low. But the thing is, people always ask us, well, what's the thing? What's the one point that says this is where you need to go start selling? And the thing is, there is no one point. It's the sum of the whole. And that's why in our real estate world, we're 93%, 94% accurate based on the feedback that

15:06
When we got into solar, we were at the 80% and then got well into the 90s because we brought in a much richer data set. Now when you say 90% like- High 90s now. Okay, so how does that number come about? Yeah, when you say our data is 99% accurate, does that mean if I knock on 100 of these doors, 99% of them are gonna buy? What's that accuracy? So the accuracy that we measure is basically

15:36
with blind tests or even open tests, right? So our customers and we ourselves are always testing our algorithms. The algorithms are getting updated very frequently, say once in two weeks, right? We have to test before we deploy because the new one that we're deploying has to be better than the previous one, right? So the way we test that is basically by showing the, or giving the algorithm known.

16:05
deals, right? Like known installs, known deals. We hold some away, we don't show it to the algorithm, and we test it against the holdout, which is what we refer to as a holdout. And how many of the holdout, meaning, when I say holdout, I mean, let's say I have a data set of about a thousand records, right? Out of the thousand records, there were 20 deals. Was Betty able to predict the 20, what percentage of that, but also...

16:31
what percentage of the thousand was it actually able to eliminate? Because Betty could very well easily say, hey, yeah, all thousand. I got all of them. But there was nothing included. So let me, cause this is, I mean, we've had two conversations now and this piece made the most sense. So the accuracy is you have this, you have Betty and you're giving Betty known data, thousand homes where you know 20 people have solar and Betty's able to tell you these.

16:59
19.99 people have solar. That's right. And eliminate at least 500 of them, right? Meaning don't bother contacting these people because these people are not gonna install solar. So it's, we're going down the product path and the output from Betty is a predictive score, right? In the solar world, we've got two scores that add up to the total score. So this is the use case, right? Someone is either.

17:28
buying opted-in leads, right? They're buying them from a wholesaler saying, I've got a thousand leads. They're acting on them. They are calling, they are reaching out, they're mailing, they're knocking, whatever they're doing on those leads, probably calling. The use case for Betty is to score those leads because not all opt-in leads are the same. Just because I might be interested in getting solar, I might be renting my house, I may be broke, right? I may not have any equity in it.

17:56
I may be going through a divorce, right? You know, all kinds of things that could preclude me. So rather than act on all a thousand, Betty can tell you, hey, there's 620, that's where you should go and reorders them. So the information comes back, it's got your predictive score on the far right-hand side, it says start here, work your way down to line 592, and then get another list because these are not gonna convert, right? If you're looking in the door knocking world, and this is what we were talking about the other day, it's the ability to soften.

18:26
Right. And one of your points that I thought was interesting, I've been thinking about all weekend and trying to come up with the formula, it's the deals, appointments set or deal ratio to miles walked, right? Or doors, miles walked. And in our case, we're saying, yeah, the door knocker will more than likely have to travel further. Rather than knocking on 100 doors, they're going to knock on 80. But rather than getting 20 appointments and five deals, they're going to get 40 appointments and 20 deals.

18:56
So their inefficiency of walking is that's where they'll lose a bit of efficiency, but they're going to get more appointments and more deals because they're going to houses that the algorithm said, yeah, this person's going to be a likely conversion for you. Okay. And just, you know, because this can be pretty complicated for us cavemen on the streets to understand. So the, back to the styles and records that you know, a certain number of them have solar, give these styles and records to Betty.

19:26
the right amount of people have already bought solar. Excluded some with bad credit scores, renters, et cetera. That additional pool of, let's call it 200 homeowners, are those houses that are still sellable? Yes. Even though Betty hasn't predicted. Yes, totally. Okay, so explain that a little bit because the way I undertook it, when I was like, well, does that mean I just want to knock on those 980 doors or why didn't Betty predict that those people could buy solar? No.

19:56
Going with your case, there's 1,000. Betty is traditionally predicting about 53%. But that'll vary because she does not grade on a curve. She's not looking at this property in comparison to the next property in the list. She's looking at the merits of that property and owner individually. So say you put 1,000 in. Betty says, these 530, here's where your deals are. Those are the ones that you knock. She's not going to tell you of those 1,000,

20:26
are your customers. Not yet. She's just saying, again, she's not telling you where the needle is, she's just saying there's no needles over here. Don't waste your time, don't do any mailers, don't knock their door, it's a waste of time. So you still gotta sell. Betty is not replacing the age old act of convincing somebody to spend money and that your product is good. It's just telling you, don't go over there, go here.

20:50
What's up guys? So proud of you guys listening, getting this far. I just love that you're investing yourself in the time, whether you're driving out to area, you're in the gym. I just want to invite you guys to take one step further and come meet us in person. We have a business bootcamp, a sales bootcamp, and a recruiting bootcamp all today, live in person events at our office where we take you, hold you hand on an intimate room, no more than 50 people.

21:16
where it's like, let's role play how to go be a better salesperson at the bootcamp. Let's role play and dive into recruiting strategies. We have 10 different modalities of how to go do campus blitzes on recruiting or how to do Indeed ads properly, or our business leadership bootcamp that is going in how to be a better manager, leader, create culture, coach reps out of slumps, training and onboarding, like all the things it takes to run a successful business. So whether you're trying to level up sales, recruiting or leadership, we have a different bootcamp for you.

21:43
and they are happening every single month right here in Salt Lake City, Utah. Guys, don't be somebody who's just, I'm here watching. Oh, I don't know if I'm worthy. I don't know if I can afford. I promise you, if you just go to the D2Dexperts.com, schedule a demo, see when our next events are, you're not going to want to miss getting involved in person hands on because that's where we really can help you. So hope to see you at one of our next boot camps. And I hope you guys continue listening and loving this podcast.

22:12
Okay, so the difference between her being able to predict which homes already have solar and which homes you should knock, like that's the part where I think people who are listening are going, okay, that's two different pieces. If Betty's 99% accurate of telling me these people have solar, right? I see, so yeah, the 1,000 and the 20 installs that the example that we took, that is only for to test betty. Right.

22:39
So you're testing her to predict who has installs. That's exactly. OK. But you can't test Betty to see who will buy. You just don't know that. Well, Betty's whole purpose is to both eliminate and also to say, OK, who would be potential buyers. Right. Yeah, and it's fundamentally an efficiency game. This is what this is about. Whether you're knocking, direct mail, phone calls, it's about making sure your resources are as efficient as possible.

23:08
and have the highest chance of success. So you're not spraying and praying, you're not knocking every single door in every neighborhood, you're being more targeted. This is what we see on our social media, on our smartphones and Netflix and TikTok, it's all machine learning based on the data that you've given those algorithms of what you want next. This is just, it's bringing digital to an analog world. Okay.

23:33
So let's talk about, as I explained to you guys on the call, at one point I was president of a solar company. We did about 300 installs a month. We were spending $30,000 a month in overseas cold callers. We had ads placed. We had canvassers, 500 leads a week from canvassers on the street. So we had all these different lead sources. And of course, my best lead source was everybody who no show to the appointments. We called back and they always schedule. Those are by far our cheapest ones. But...

24:02
Now getting these leads into the proper way to use Betty, right? Randomly, the call I had before I came in with you is somebody who was a virtual salesperson for Solar, for one of the major players in the space, and where he never would knock the door. He just closed leads that were put on his calendar. Have any of your customers used Betty to target their...

24:31
The meeting I had earlier today, by the way, was a company that geofence advertises for recruiting over college campuses, right? Same kind of geofence advertising over Betty's homes, right? To generate possible leads and or any kind of old school telemarketing kind of approach. Has anybody used these methods so they don't? Because I'm just picturing, if you overlaid Betty's data onto a

25:01
a map that I'm knocking. Now, if it's 55% of the homes in my neighborhood or somebody I should knock, okay, great. But if it's 22%, then I'm like, oh my God, that's two out of every 10 homes and I might have to knock, walk five blocks before I knock another door. And that would be, for a business owner, counterproductive to arm our salespeople with that data on a kind of a knocking app. So, kind of a two-part question. Has anybody used Betty's data outside of door knocking?

25:29
And then how have people used it if they are knocking? So the easiest way from a non-technical standpoint, like me and you, Betty fits in right after raw data, right? Whether you're buying that data or you're getting the data, you're buying a lead, that's when it goes through Betty. And then that information is overlaid for door knocker, just what you said. We've had customers who say,

25:59
I'm not putting my resources on the street there. I'm gonna go over here. This is 70% is where Betty thinks there's an opportunity. So instead of putting my resources here, I'll do direct mail there, right? So Betty doesn't replace any of this outreach. It just makes sure whether you're cold calling, direct mailing or knocking, that you can be more tactical. But it fits in right after the data and then before your action. And it gives...

26:26
owners and directors of sales and door knockers, insight and intelligence about what they're about to do. So rather than blindly running into the wall, one of the things that you guys do is you're teaching people how to do it. This is the opportunity for them to be more effective because they're gonna go in with a higher probability of success. So it's after data collection and before action. Once Betty has done that, then it gives the owner or the director or whoever, hey, whoa, I don't wanna do that. That's not good.

26:56
So it's just a matter of getting intelligence in advance of your action. And then how to use that intelligence, right? So you send half as many mailers, right? And you just call half as many people, so your people on the phone are being more efficient with their hour and ultimately knock half as many doors. And one of those, is there a leading candidate of which way this data is being used the best? Because I-

27:26
I could see, wow, if I was running a call center and I don't know where I'm getting my leads, but if I am, can I use this data to generate leads through other traditional sources? Because it's kind of like a lookalike audience on social media, but it's different when you're using the data you are. But how do we use that? If I ran a call center, how would I use Betty's data to call better lists? So you could replace your current data supplier and say to Betty, hey, I need.

27:55
20,000 in these four counties, right? Give me 20,000 high probabilities. And then she's gonna give you 20,000 high probabilities. Not based on your data, but based on the millions of lines that she's already run. She's gonna go out, do her thing, and come back and give you 20,000. And then if they have phone numbers, you have that, and you throw them to your dialer. You still need to skip trace it, right? Betty does not skip trace. We always encourage people doing calling, get a skip trace by a top notch, right? So you keep out of trouble. If you're a door knocker.

28:24
and you're in a CRM that provides a canvassing app. So, okay, I want to look at these neighborhoods and you've got all the, you're able to get all the households, right? You take that data and you run it through Betty and then you can get your Betty score and you can overlay it on your map and say exactly like you said, that's not a good neighborhood. I'm gonna go another, I'm gonna mail in that neighborhood and let them call me rather than having to put my resources out there. There is no one way to do it.

28:53
The most important thing we want people to take away is it gives you intelligence and it gives you insight with about what you're about to do. As a business owner or a head of a sales department, you can make smart decisions based on your own history because you're getting advanced knowledge of what your team, whether you're calling or knocking or mailing. In the real estate world, it saves people about half of their mailers. 65% of our customers in the real estate world, they use it to scale. They say, no, my budget's...

29:23
$15,000 a month in mailers. I'm gonna keep that budget. So now instead of 15,000 raw mailers I'm sending out, or my previous scrubbing, I'm gonna get 30,000 lines of data and get down to 15,000. So if I was getting 100 properties based on my 15,000, now I'm expecting to get 193. Same marketing expense, same cost, I'm just getting a better product out the door to a better target so I can get.

29:52
double the deal. So about 65% in the real estate world use it to scale. So I'm like, no, I'm going to lower my cost to keep my deal flow the same. The bottom line is there's no one prescribed way to do it. It's just a matter of saying this machine learning can allow people to make smarter decisions. Okay. And now you just launched the solar, right? So in the solar space, are there champions that you're working with that are using it a certain way? Like the, is it the door knocker? Is it the mailer? Is it the, is it the,

30:21
call center, how has it been best practice? Yeah, so we've been live for a couple months now. We've had, we started with beta customers, right? Just to, hey, try it out. Let's get some real world data. Let's take it out of the laboratory, give it to people and have them run their data through it. We let them run historical tests, you know, because you got to convince, run a historical test. Wow, it works. Okay, now use live. It's primarily knockers, right? But it's prevalent.

30:51
in the solar world. And so we're about 80% of our champions are knockers. We've got a couple that are using it to score opt-in leads, primarily because they were frustrated with the quality of the opt-in. And their call centers are not performing as well as they used to. They don't know if it's because the market's getting tough or because their opt-ins are not very good. And so now they score it.

31:21
and they are able to cut down on their callers or allow them to be more efficient. And so it's about 80% callers or knockers and 20% callers. So if I'm a business owner and I'm listening to this, every business owner in the world is like, well, how much does this cost? And obviously you guys are smart enough not to just drop prices on a podcast. You want them to reach out to you. Actually, we can. We can. And this is what gives me goose pimples. There's no people behind many.

31:51
Right? When somebody is uploading a file or it's going via an API, we don't have somebody sitting offshore, clicking on the file, dropping it over here, doing some work on it, getting it back to you in three days. No, it is all automated. Like 99.99% of the files go through. You can see where your file is progressing. I mean, just to give you a perspective, we run five to six million records a month, you know, on a daily basis, weekends, especially also. Weekends are big. And...

32:20
We've been able to support, if I share that, you'd probably laugh, how many people we have on support team, but like very like in the onesies, twosies, right? Because we've been able to scale the technology. Because you built it well. So the pricing, it runs from 45 cents a line to 20 cents a line, right? Depending on how much you want to run through it. There's no commitment. A line is? A house, a structure, a person. Okay.

32:47
So if you've got 10,000, you're going to be in the 30 to 33 cents. It's our- Was that, is that the math? Is that $330? $330. Okay, $330. Right? And that should be able to, historical averages, cut out half of your activity. So some of the larger door-to-door companies will go to different markets. Some companies live in an area. But the summer model-

33:16
you know, ABC Pest Control Company or XYZ Alarm Company is gonna go to Boise, Idaho, and they're gonna knock Boise, Idaho for the summer. And Boise, Idaho, I don't know, population half a million people, right? And so you're probably half a million, you're probably in the 20 cents range? Yes. Okay, so you're paying 10 grand to, my math is right, 20 cents times half a million. You're paying $10,000 to scrub all of Boise, Idaho to give you...

33:45
to take away about half the doors that you shouldn't knock. And if the technology is probably not there yet, but eventually, I think it would be really smart if for you guys to be able to find a way to API into some of the major canvassing softwares. And many of these companies have their own canvassing software, where you open up the app and all the red pins are their current customers, right? And so if the blue pins were Betty's customers, right? Then that even sweetens the.

34:13
So we realized early on that APIs were a crucial part of our customer's business. And so we built APIs, right? We have not built an integration with a canvassing app yet. It's something that's in the works for a couple that were, because it's important, right? You get the information, you run it through Betty, and you're putting it back in your canvassing app. There's some great canvassing apps out there that are really wonderful tools, right? And...

34:41
we're looking to be inside of it so people can just say, okay, I'm gonna still use my canvassing app. I'm gonna run it through Betty to make sure it's all square. And this is where I'm gonna go and I'm gonna have a visual display for my knockers to be able to do the thing. Yeah, that is two, three months away. Yeah, and our niche and our focus is purely what we're doing, right? Which is machine learning and lead qualification. And yeah, the volume that I mentioned earlier, the...

35:10
five to six million. A lot of it is actually through existing CRM integrations on the real estate side. So we are already well integrated on that front. And we did start from a sort of a B to small b model, but it's very clear, got very clear very soon on the real estate side that, yeah, this way to scale is actually, we're a data provider company. So we got to integrate with larger companies. So, and yeah, we are in the works with at least two or three companies on this canvassing side. And yeah, we hope to.

35:38
Yeah, and one of the points that Sean Mayberry made there, JP, is the machine learning is our wheelhouse. As we initially were putting this together, we were thinking, oh, we could resell data. We could get data from another provider. We could build a CRM. We could skip trace. We kept coming back to do one thing and do it really well. We're not going to get in your wheelhouse. We're not going to get in your wheelhouse, but we can integrate. You don't have to recreate it. We've spent a lot of money.

36:08
on data acquisition and augmentation to build, it's the pharmaceuticals, right? That first pill is a billion dollars, the second bill is one cent, right? The training and the building of the algorithm to get it where it is was very expensive. Running that now, it gets cheaper and that's why we're comfortable talking about prices because we're in the sense because it's our cost, our major cost is computing power and augmenting data.

36:36
And we'd be very open to, of course, white labeling and the larger company, the canvassing company might sort of have it as a premium feature on their platform. And yeah, we could call it Betty or call it whatever they want. So just to recap for some people that are listening, it's a big understand. There was also Blitz models in this industry, especially in solar, where you live here in Utah and you're gonna travel out to...

37:04
Albuquerque, New Mexico to go blitz solar. And we assign territories through these knocking apps. Hey, you're gonna have this area, you're gonna have this area, you're gonna, and you can get a really good idea as a leader, hey, we're gonna knock this area. And we even know kind of the record count from the knocking app. Hey, there's 27,000 homes in this area that we're gonna knock in this blitz. Okay, great. I have where I'm gonna knock. I have it mapped out.

37:33
How do I get that area to you so you can run Betty's data? Because honestly, 30,000 people and it's gonna cost me 1,200 bucks at 40 cents. It's nothing. I'm like, and I'm bringing a blitz team where I'm gonna go sell 50 solar deals in 10 days and I'm gonna sell 500 kilo. It's nothing to spend 1,200 bucks to clean up my data.

38:01
So I would put this as part of my Blitztrip planning. So give me the timeline of what would need to happen if I knew where I'm gonna knock and I can roughly give you the, I'm not gonna do all of Albuquerque, but I'm gonna do these 12 neighborhoods or whatever. What's the turnaround time? About, I was, depending on the size of the file, say it's 27,000, six, seven hours. Okay, and I actually have to give you the file? Yeah, so I just wanna jump in on that. You could do two ways.

38:30
A lot of these canvassing apps have features, you know, where they have the owner name and address. That'll be ideal if you could, you know, if there's an export of the owner name and address, you feed that into Betty and, you know, Betty will take, you know, for a hundred thousand record, might take four hours or something like that. What if I don't have a knocking app? What if I'm a small company, I don't use one, I just show up and Albuquerque? Sure, or you could also go to our site and there's a lead gen feature, which uses some of the exact same algorithms.

39:00
and you'd put in a zip code or a county. Or multiple zip codes. And then you'd get the same list as you would with the other one. Okay, so I could go, I could give you, here's the seven zip codes I'm gonna knock on Albuquerque. I can put them in there. And then the next morning, you charge my card. I have a filter list. I would probably, you eventually would probably have to find some simple solution for door knockers.

39:28
to some free mapping software, right? So it's very- Or we integrate. Or you integrate, right? And I'm just thinking, people listening to this, we're an impatient breed, right? Like there's somebody who's probably listening to this- By nature. Right, they're probably already Googling, right? How do I get a hold of these guys? Who are these guys? Because for that small amount of money to get a competitive edge, people are interested in, and so they'd be like, okay, now I'm using one of the larger canvassing apps. They're not gonna give me the data.

39:57
to run through it. So I'm gonna have to give them the zip codes. I'm gonna have, you know, and so what I would do is I would find some free mapping software and I would overlay those points, you know, through an upload and then, and then what, and so obviously if you guys had a free mapping software that you could link to somebody. We could build a case study. Yeah, and also the people that are, if they're using one of the larger canvassing apps, right? They can have a conversation with them about, hey, I want to integrate with this.

40:25
You guys happen to come on a good day, because I think two of them are here. So I can make some introductions for you. Appreciate that. We have our sales boot camp and our recruiting boot camp happening this week. So you guys happen to come in. I saw one of them setting up before. And so I can make some introductions. I actually had a call with another one earlier today. So there could be a good synergy there as well. All right, so how do they get a hold of you? So rework.ai. My email is Fred at.

40:53
rework.ai. Shams email is sham, which is S-H-Y-A-M at rework.ai or admin at rework.ai. You can go to our website. There's a button to book a call. We get on a Google Meet and chat through your personal circumstances of what you want to do. That's where we get a lot of our feedback. We love those calls because it helps us learn about what people are going through. We'll also be at D2D Con. We're really looking forward to that.

41:24
And, you know, it's, yeah. That's the best part. It's designed to be fully self-serve, but as Fred said, we'd love to, you know, connect with you before you use it, if that's a thing. So, yeah. Yeah, great. Well, again, hey, door-to-door experts here with Reworked AI, Fred and Sean, appreciate you guys coming in. And I think those of you who got to this part of the podcast are pretty mesmerized. We'll talk to you soon. Exactly, exactly. Take care, everybody. We'll catch you next week.