Alex Boyd: Just by doing the estimate in person, stopping the recording, boom, estimate's done. I can literally send the full scope estimate while I'm still talking to the customer. AI doesn't forget. It'll take all of that and see that in your transcript and when I'm doing these quotes. Every single time I'm like amazed. Oh my gosh, I completely forgot that the customer said that. I completely forgot that they mentioned that. It's better than me. 

Austin Gray: Hey, welcome back to another episode of the Ownr Ops podcast. I'm your host, Austin Gray. In this episode, I have Alex Boyd from Brushwork Land Clearing Services, joining me again because he just put together this incredible AI automation that automates estimating for him in his land clearing business.

Alex Boyd: Like the idea of automating your sales process from a little pen. From beginning to end without ever taking a note or without ever looking at a computer screen For a sales rep like that seems crazy to most people right there. No way. I can't do something like that. 

Austin Gray: So by the end of this, you are going to learn how he uses a pendant to record his conversations, transcribe the conversations, summarize them, send 'em over to Chad, GBT, and then ultimately create estimates that gets sent over to his estimating software. This one was so much fun. I can't wait for you guys to listen to this. 

Alex Boyd: At the end of the day, people don't buy what you do. They buy why you do it. A company needs to have an identity. It needs to have a personality. 

Austin Gray: So one framework I'll share, if you don't mind, we had an AI expert on the show months ago, but he said, use this as a general framework anytime you're prompting chatgpt. But real quick, if you have not already, we have a free school group specifically for blue collar business owners. You can join for free skool.com/ownrops skooll.com/ownrops. I share things like how to go optimize Google business profile, how we're running profitable Facebook ads, and how to set up and optimize jobber so that you can save time on your back office admin tasks. Without further ado, let's jump into the episode. This one's gonna be fun. Alex with Brushwork, he does nothing and estimates get created with ai. Welcome back to the show, Alex.

Alex Boyd: Thanks. Glad to be on.

Austin Gray: So you just did a video, YouTube video on creating estimates using ai. You have a pendant that records you while you're on site with the customer, and you have custom gpt set up to calculate your pricing and then send it to your estimating software you were just telling me before we jumped on the show. You even have it set up further than what you showed in that video. Can you explain a little bit more?  

Alex Boyd: Yeah, so basically I've got it to the point where can just stop the recording. It will upload the recording into plot, which then sends it over to Google Drive, which then is pulled   off of drive into A GPT and breaks  down. Like what I showed on the video where it says the full summary of the conversation. Then it'll take that summary, send it to another GPT, and it will then use our formula for pricing, our line items and descriptions, and then it'll take that, parse it into JSON into a house called pro API, to where it'll automatically create the estimate with accurate line items, prices, and descriptions along with our operator notes  for the job and send   that to the customer. So I can essentially just by doing the estimate in person, the recording, boom, estimate's done. I can literally send a full scope estimate by just while I'm still talking to the customer.  

‍Austin Gray: Absolutely incredible. This is a way that AI can help you rapidly grow or at least save time in your service-based business.

Alex Boyd: Oh yeah, the amount of time it saves. So think about the times.  If you've ever had somebody come to your house they come out and they give you an estimate, a lot of times they'll just sit in a driveway for 30, 45 minutes while they compile this estimate together, and then send you the estimate, come back in, present you the estimate, why you don't need all that. Just you can cut all that out with ai. So one person can now do two to three times the amount of estimates in a day, and the estimates are better because AI can pick up things that, conversations sometimes that you miss, and compile all of that and it takes better notes. It's more detailed. You just have to train it on exactly what you needed to put in.  

Austin Gray: How accurate are the estimates? 

Alex Boyd: More accurate than me. So what I mean by that is if I'm going out there and I'm putting the notes in myself,  and I'm going off memory, 'cause you don't wanna be rude, you don't wanna pull out your phone while you're talking to the customer, right? And just take notes as you're walking through. So you will have to go by memory and you go get in the car and you're like, okay, everything's fresh in your mind. And take the notes on there. You forget things, right? So you'll, you might forget that there was a big steel piece, a big piece of steel somewhere  buried in the woods that the customer showed you when you go  out to create the estimate or you need to create the operator notes. AI doesn't forget. So it'll take all of that and see that in your transcript, and it puts it in there. And when I'm doing these quotes, I'm, every single time I'm like amazed oh my gosh, I completely forgot that the customer said that. I completely forgot that they mentioned that. So it's better than me at doing it. And it's more accurate. So I think that's, that's where AI is now.

Austin Gray: Everybody's biggest question right now, especially if you're in the blue collar industry. Do I have to ask permission to record these conversations?  

Alex Boyd: It depends on your state. Some states are two party consent the plot note is HIPAA certified.  I'm not in the medical industry, but I know the HIPAA certified is like the highest you can get on privacy. I say something to the customer just so they know, I'll say in Ohio we're my understanding of one party consent, but I just will say, Hey I'm gonna record the estimate just so that I've got better notes at the end of the day. Is that okay with you? Nobody ever says no. Because they want it to be accurate too, right? They want 

‍Austin Gray: Absolutely. Absolutely. And as long as it's framed as. I'm using this as a tool so that I can be a hundred percent present in the estimating process and really deeply listen to the conversation rather than me being distracted. Like everybody's, you've, we've all dealt with that person who showed up and yes, while you appreciate they take notes, sometimes it doesn't feel like that person is actually listening if they're just looking at their phone. And I believe that, some, a lot of the customers we sell to are 60, 65, sometimes 70 years old.  And if you're pulling out your phone or an iPad, sometimes it can rub them the wrong way. And so this is why I've been super interested. We use the the, it's called a limitless pendant. Same deal. I have not dove into the research. It's pla, P-L-A-U-D, is that correct? 

‍Alex Boyd: Yeah. Applaud. Yep.

Austin Gray:  I recently got back from launching a land clearing business down in Austin, and this last winter I launched a snow shoveling business alongside Bear Claw. In both businesses, I've implemented jobber as a way for us to efficiently manage quoting job schedules and invoicing, and even collecting online payment. Why? Because it's worked so well for us in Bearclaw and it's saved us a ton of time and headache. So if you are looking for a software that can help you manage the back end of your business, look no further than Jobber, you can visit Go dot get jobber.com/owner ops O-W-N-R-O-P-S.

Austin Gray: So same type of concept, I assume it just records the conversation. Correct. And then transcribes that into a piece of software. Is that right?

Alex Boyd: Correct. Yep. Records the conversation, transcribes it in the software, and then use AI to, generate the summary. So whatever you're doing, so you give it the context and the, and what you needed to generate. Same as making a custom GPT. It's very simple. And then it'll use that to create the summary of the conversation. So any small talk, anything that's not relevant we'll get filtered out with your or with your prompt.  

Austin Gray: Fantastic. So for anybody listening to this on the OWNR OPS podcast right now, whether it's on my YouTube channel, whether it's on Spotify, whether it's on Apple, or whether you're watching this inside the school community, Alex  has already created a YouTube video where he walks you through step by step how he has created this and also how the process flows. So I don't want to dive into exact screen share on this episode. I want to talk more high level here on actually how the process works and even dive deeper into your thought process behind how the actual automation works. So if you want the step-by-step process, we'll put the link in the description here and you guys can go watch that, how to video from start to finish   where he shows you how to implement this. Let's keep going  though. On the pendant records it puts it into the software which comes with the pendant, and that is connected to ai has its own 

Alex Boyd: yep.  

Austin Gray: AI embedded in that so that it automatically summarize it. Do you have a prompt inside of the plot software that you have had to train for it to summarize on?

‍Alex Boyd: Yeah. Yeah. So we we created my own custom prompt and it's very simple to set up. It's basically they give you a sample of what the prompt is looking for. just copied and pasted the sample and to chat GPT and said, Hey, I need this to do this. This is what I'm doing. I'm going out and giving quotes. This is who I am  this is the things that I'm looking for. And then GPT then spit that prompt back out to me and I copied and pasted it back over to plot. So it's super easy to set up. Anybody can do it and they can adapt it to anything you're doing. If you've got a plumbing service, you're an electrician, you're a painting company, a power washing company you can do this with anything. And then it created that kind of perfect prompt that creates a summary 

‍Austin Gray: So you're using chat, GPT. To ask it to create a prompt for plot 

‍Alex Boyd: Correct. Yep.

Austin Gray: Now your prompt frameworks, everybody has their own tweaks on how to prompt GPT correctly or their LLM correctly. You like prompting?  

Alex Boyd: I like to be very direct, I just say if I'm, if I paste that sample in, I'm just gonna say, this is what we're doing, this is what I need. This is just be very direct on my prompting with GPT. that way if, it gives me fluff or something on the way back, or if it's trying to use too much natural language, I just want the result, I'll just reply only, give me the result, because fluff wastes time. So I'm pretty sure I've trained my GPT at this point to just constantly give me super direct answers.  

Austin Gray: Yeah, absolutely. Okay. So one framework I'll share if  you don't mind. We had an AI expert on the show months ago, but he said, use this as a general framework. Anytime you're ch prompting chat, GPT, hey, chat, GPT, you are an expert in whatever you're trying to do. Tell it more details, and then at the very end of it, ask me any clarifying questions that you need from me in order to do this to the best of your ability.

Alex Boyd: Yeah, I do that for custom GPTs. 

Austin Gray: Okay? Sweet. 

Alex Boyd: But if I'm just interacting with the, the regular dashboard I don't do it as much. But for a custom GPT, like the one that's trained on our pricing model, that's exactly what I do. So you are a brushwork pricing expert, and then, ' cause it does give better results.

Austin Gray: Let's dive into that now. I'm glad you brought it up because I want to hear your thought process behind how you went and trained your custom GPT on your pricing model.

Alex Boyd: So basically, again, I used. Chat GPT to help me make a custom GPT, right? So in that case, if I'm making cus a custom GPT, I'll be like, you are a custom GPT expert. And then I'll just brain dump or dump in my entire pricing model. The  price per quarter acre, our minimums, our discount rules, every single possible thing that's in your pricing, you can put into that model any qualifiers, like if it's over this amount of density if  it's over this size, all of that stuff you can just dump in into there and then have Chat GPT, make your GPT right? And then after you do that you put it into a custom GPT, and then you just test it. So you start putting in the prompts this is what, it's, this is what I needed to  put in. Is it giving me the right result? And if it's not giving you the right result, it back into the regular GPT and say this isn't giving the result I want. This is the result that I want. And then after a few iterations of that you'll get the perfect result every single time. that's how we trained our pricing.

Austin Gray: So in your business, your land clearing forestry mulching specific, you're using a skidsteer with a forestry mulcher. For listeners who don't know Alex or his business, he's using a skidsteer with a forestry mulcher and he is charging people to come clear underbrush. We have a very similar business out in the western United States and Colorado. We use similar machines, but it's a little bit different and we reduce vegetation in the forest for wildfire risk. One of the challenges that we have in our business, and I'm sure you had this early on, and I'm curious to hear how you have put your pricing model together, is how do you calculate your per acre cost of clearing land based on density?  

Alex Boyd: So that was a challenge and we started with the per day pricing. And the per day pricing was easy, right? But when we're doing that the per day pricing has problems in that. always a winner and a loser on each side. So if you tell somebody it's gonna take a day and you only get half their project done, and you're going back to them and saying, Hey, please give me another $2,000 for another day. And the customers  don't like that. But while we were doing this pricing  model, we were logging all of the machine hours, we kept track of and then we would take before pictures as well. So we'd have pictures, the machine hours, and then we went back to create the. So like we went back to each job and we a hundred plus jobs and just out, okay, based on this terrain and what this  brush looked like, it took us this many hours of running. And then we averaged that We basically ranked it in three different densities of brush and three different levels of terrain. 'cause the harder the terrain is, the longer it takes, the dense of the brushes, the longer it takes, right? that's essentially how we came up with that, our terrain and density model. And  obviously we had Chat GPT help us create this the whole time. So it wasn't just like we were crunching the numbers to, as on our own. just provided it the data to help us create this. And now we quote by per project price. So that way, the customers are happy, they get their project done and there's no nuance or there's if we quote a project price, we're very confident that's an accurate price for us and for the customer.

‍Austin Gray: How has the development of that pricing model been from the time that you started training GPT? Let me ask, let me frame this question a little bit better. I'm curious about the transition from per day rate to per project, and were there any projects that you missed out on from the company side? 

Alex Boyd: You mean where the customers didn't approve the project?  

Austin Gray: So you were charging per day, now you charge per project but the only way that you have the confidence now to charge per project was to go through this testing phase of creating this pricing model  with chat GPT and then testing  that to see if you could actually come in. For that price, for the time that you allocated for it internally, were there any jobs that you guys either broke even or, lost money on during that process? Or was it pretty straightforward? 

Alex Boyd: It was pretty straightforward really. We had a pretty good idea of what was profitable on the per day side of things. So I quoted a project on a per day basis, I said, Hey, this is gonna be two days. I've got a pretty good idea in my head around where I, that project pricing should land and so we, at the same time we implemented project  pricing, we increased the price of our service as well. And so when it spits out a project price, I would essentially compare it to what I would've quoted per day. And if it comes out on the higher end, then I'm pretty happy with it. So we haven't necessarily lost money on the project pricing, but we absolutely lost money on the per day pricing. 

Austin Gray: How So on per day pricing. 

Alex Boyd: So on the per day pricing, what would happen is, let's say you're quoting the, you're saying one day it's, you're gonna run the machine for eight hours. You get 90% of the job done, of the project that they want completed. Done. You're like, oh man, I'm just gonna push a little bit further. We'll stay a little extra time, we'll get the other 10% done. We're not gonna charge the customer. 'cause it's hard to go back to a customer and ask for more money, right? So we're not gonna charge the customer the extra money. so then your operator would be out there for another two, three hours just to get their project all the way completed. And that's just time that you weren't getting paid for. You feel obligated to  do it. You don't want a bad review. You want the customer to be happy,  right? Over time, that extra time that you're spending starts to compound.

Austin Gray: Stryker Digital specializes in SEO services specifically for local service businesses. Bodie and Andy, the two co-founders, have helped me get Bearclaw Land Services to the number one search result on Google inside my state for my specific search term. If you wanna learn more, visit strykerdigital.com. That's S-T-R-Y-K-E-R digital.com.

 Alex Boyd: Or there's in another case where you might have quoted two days and you've allotted for two days on your schedule for the operator to do that job, but it only took them one day. So now you're refunding or you're only  charging the customer for one day, but you've also lost a day on the schedule that you had planned on the machine being out there. So you're constantly losing out on opportunities. So you have to switch to a better model. But the problem with a better model is it takes data to do it and to do it well .

Austin Gray: Takes data or an experienced estimator who understands. 

Alex Boyd: Yeah.

Austin Gray: Because that's where the risk is either in your head as the owner or in your, one of your key team members heads as all of our estimating processes in his head or his or her head. Because,   they've estimated this   many days on a job. They understand our current crew and the rate that they can move through. They have this gauge in their mind, like they're subconsciously gauging that, how dense is this forest? Or how dense is this land? How steep is it? And they're. Placing their best guess for how many days it's gonna take the crew. But then if that person leaves that job, or if decide that it's time to hire, how do you train the estimating process for the land clearing or fire mitigation industry without a pricing model like you've created? And so the reason I'm bringing this up is this is a topic I'm super interested in, and I think you've done a really good job of creating a process that can be   repeatable. And so for our listeners here I really want to encourage you guys to go listen and watch the video that Alex created. I watched it last night and I watched it again this morning. But it, this is one way that I believe local service business owners can and should be using AI to be more efficient. 

Alex Boyd: Yep. Oh, a hundred percent. Yeah. And I feel like, the, this local services that have been around for a long time, if they don't do this, they're gonna get left behind because the people who do are gonna have half the sales team produce double the quotes.

Austin Gray: Yeah, absolutely. All right, let's talk the, let's talk your pricing model and how Chat GPT help you break this down for the land clearing industry. You said you have three different tiers of vegetation density and you also have three different tiers of steepness of grade on the land. Is that, did I hear that right?

Alex Boyd: Yep. So initially we started, it had given us five different tiers of brush density, but that was too complex. And so I said, Hey, I only want three, because three makes sense to me because there's two different ways to look at brush density. If you're experienced, you can look at brush density and say okay, this is  really thick. I can hardly walk through this, or, this opens  up here, and you can gauge it. However, there's another way to look at brush density and say especially if it's. A common area, like if we're only in southwest Ohio, the woods look pretty similar, right? So the other way to look at density is to say, okay, this, they want this to look like a managed forest. What that means is we're gonna be keeping small saplings.  How native saplings, we're basically only  clearing the underbrush. It's still gonna be a very forested look. And then another way to look at tier two is that they want this to look like a park. So we're gonna keep these four to five inch trees. And it's gonna be mostly mature trees left the saplings and the brush will be gone. And then the tier three is gonna be, they want everything possible the machine can take down. So you're  only gonna be left with large trees. Because as the larger the tree, the more time it takes to process it. So that's how I look at tier three more for, if I'm training somebody on how to do this, is the, that's the easiest way to look at it. And it's the thing that the customer understands. 'cause I can say to them, do you want this to look like a managed forest? Do you want it to look like a park? Or do you want as much as possible? And the customer understands that. 

Austin Gray: So your questions, you ask them, do you want this to look like a managed forest, which would be AKA.  The lowest amount of vegetation that you would be removing in that scenario as an operator, do you want this to look like a park, which would be, you are removing all underbrush in tier one and some of the. saplings, or what was your, what's your cutoff point for.  

Alex Boyd: Yeah, pretty much all of the small saplings, trees at around four inches in diameter, three to four would stay. And typically, unless they're in like a cluster. So if there's like a cluster of three or  four where maybe only one's gonna start 'cause there's gonna choke each other out maybe we'll take, we'll leave one of those. And that's more of a park setting. It's still very open, but you still have a good amount of healthy trees. And then, yeah, the tier three is just whatever the machine can take.

Austin Gray: Where do you put the cutoff though, for that? Because theoretically you could go right into like most any and I don't know the size. You may have absolute giants out there, but. You could go mulch right through a tree and then spend, hours mulching that tree up theoretically. 

Alex Boyd: So our cutoff is six to eight inches depending on the species. So if it's a hard, we have a, most of our trees out here are hardwoods. We have some evergreens which are pretty soft and we can go like up to eight inches in diameter. But of our hardwoods will stop around six inches. And the reason we stop there is not because we can't take it, it just takes so much time that it's not worth the customer's money, and it's not worth the strain on the machine. 

Austin Gray: Okay. Okay. 

Alex Boyd: These hardwoods are hard on the machines soft woods are a lot easier. 

Austin Gray: No doubt about it. Now this for listeners, if you're in land clearing or forestry space, our stuff out here in the west is gonna be completely different. I'm trying to take what Alex is doing by his tiered approach and think how this would relate to our forest out here. And I'm going to have to put some more thought into this for how we would tier it. And I'll use Chat GPT after this to work through it. Same deal. If you're down in Texas, like complete different landscape you're mainly just removing one type of vegetation, which is cedars. And cedars can be super dense. Other cedars can be small and not that dense. And so I think what I'm taking from this, Alex, the framework that you've created for land clearing that you found to work is a three  tiered approach. Hey, tier one customer, do you want us to remove? Light underbrush tier two. Do you want us to thin this out and make it look like a park? And then tier three, do you want us to come in and remove as much as we possibly can?

This episode is brought to you by Dialed in bookkeeping. Ben and his team provide bookkeeping services, job costing reports, and accurate financial information for the home services industry. If you're looking to keep your books up to date, visit dialed in bookkeeping.com/owr ops. When you use this specific landing page, you'll get your first three months, 50%.

‍Alex Boyd: Yep. That's it.

Austin Gray: Cool. How  about steepness of grade?

Alex Boyd: So for terrain it's more straightforward. It's a, for something that's flat, it's obviously gonna be your tier one. Mostly flat. Basically the skid steer can handle it in any direction. It doesn't matter which direction we hit this tree or this brush at the skid is not gonna be sliding. Tier two is a little bit steeper grades, so around 10 to 15%. It just depends. usually that's gonna be, Hey, I've only got. I have to come down to the bottom and hit it up, or I can only hit it from this side. It's just more time that the operator's gonna spend maneuvering around to, to get that area. And you could break it down by exact grades. I just haven't really done that. And then tier   three is gonna be, there is only one safe angle to hit this at. So either that's gonna be coming up from coming down from the bottom and working your way up, or maybe there's only one little access point to where it's just that steep to where if we, it's too steep somewhere else to run the skid. And we run the skids up to about a 20% grade mainly because we have, we have slippery soil. It's very wet out here, right? And the heads on the machines are heavy, so we don't want to be on too steep of a terrain, even though I think the machines are rated like 25 or 27%. We really cap out at 20 before we will, the green climber. So beyond tier three is green climber, which is a whole nother model of pricing. But that's the way that I look at terrain.  

Austin Gray: Now, can you share, let's go back to your day rate pricing. What was the day rate pricing you were running with?

Alex Boyd: So we were at 2000 a day plus the $200 mobilization. So it was 2200 to run it for a day. We didn't start with that. We actually started at 1800, 1600 so 1800 for the first day, 1600 per subsequent day. we moved up and we did it intentionally. More of a, when you start out, you need momentum. Quote low quote low is gonna get you momentum, right? you're not gonna make much money, but you, it's going to, it's gonna give you momentum. And then you can incrementally increase your pricing over time. And then once we hit that 2000 a day, that's when we said, okay, we've gotta start looking at this from a different angle. We've gotta start doing project because now we have data where we can base our project pricing off of.

Austin Gray: Okay. And once you move to the project pricing and you've tested this, where have you landed on a tiered scale? 

Alex Boyd: What do you mean? 

Austin Gray: Does it kick out a per acre cost?

Alex Boyd: Yeah, so the scaling is basically by quarter acre. and. And the reason is because typically, on that small of a bite, I can say this quarter acre is a tier three, tier two, right? Tier two density or something. Tier three terrain. So you need small enough bites to separate the property out as you, I use LandGlide to highlight the areas and to measure it. So one property can have multiple densities and terrains.

Austin Gray: I haven't used LandGlide yet. We use OnX maps 'cause I'm a, I'm an elk hunter and so that was just the mapping tool that we had access to when I started the business. LandGlide. Do you mind to show us how you use that?

‍Alex Boyd: Yeah, so LandGlide. The reason we love LandGlide for us and land clearing is the clear property lines. Other apps don't really show big, bold property lines, but LandGlide does. It's very simple to use. So I can train a operator to use it. Now sales uses it, but we also have, the operators will  use it. Um, so when they're out on a job site, they need to be able to see where the property lines are. We have used OnX before, and that's typically gonna be for if we're cutting trails and the customer wants an exact for the trail. 

Austin Gray: Okay. 

Alex Boyd: Because on OnX we can map the trail and then share that with the operator, who then can pull up OnX and follow those coordinates. So we do use OnX. It's just that it's for specific use cases, whereas LandGlide, find it's easier for property lines and for measurements.  

Austin Gray: Easier for property lines and for measurements. Okay, so you go into LandGlide and you highlight a quarter acre area, and then how do you, how does this integrate into your pricing calculator?

Alex Boyd: If it's separated, most of the time it's not, most of the time it's, it's a tier one, tier two it's consistent across the whole property. But if it isn't, then I'll measure the different areas and then I'll just feed that into the GPT pricing model. This area is, tier one, tier two, this area is. Tier two, tier three and then the GPT will then spit out the correct price based on that area. it'll take both of those as separate pricing then combine them together. It'll, it'll  create, I can create two different line items too. I can do it just depends on whatever I tell it. I want, like I can say, give me two different line items and then it'll give me a line item with mild terrain, medium brush, and a line item with medium terrain, dense brush.

Austin Gray: Now have you found that your tiers are accepted well by your customers on your estimates?  

Alex Boyd: The tiers. So internally we use tier one through three. What the customers see is mild, medium dense, or mild, medium, steep. Because on an internal level it's easy to just say, give it a number, but on the customer's side it spits out, something a little bit more descriptive that they understand. So far we haven't really been asked any questions about what it means. And I trained that our GPT to do that. So if the input might be a tier one, but the output should always be this, like the out output should be mild can be tier one, it gives it a little bit more of a descriptor. And then I also have on the description, it'll take the part of the scope of the work that pertains to that. So if it's forestry, mulching. And then, and the customer said I was out doing the quote, they want us to do this area along the creek, or they want us to do this area that they can, that's only malleable. In many cases that happens. It's I only want you to cut down what I could actually hit with a Mount earth and nothing too steep. That GPT will automatically add that into the description and then of the line item. So it's very descriptive for them. They have a general idea. You don't necessarily wanna break down your pricing exactly. You still have your internal methods but you gotta give them enough of an idea and a description that they understand what they're looking at, that it's not just pulled out of thin air.

Austin Gray: Do you include? Screenshots of the maps or does your automation pull the areas from LandGlide, for example, if you have a you have a mild density area that's two acres with medium grade, and then you have another area that's another two acres with dense forest and steep grade. Does your automation take a picture on LandGlide or a screenshot and add that to the estimate? 

Alex Boyd: No, not, that part's not automated yet. I haven't quite figured that one out yet. mainly because you have to go in and mark the area. I think, if it was like we're doing this whole property from property line to property line and then LandGlides got, this property is 56.35 acres or something, then you could automate it. But because it has to be, have to mark the exact areas you're doing you still need human input there. 

Austin Gray: Sure. Okay. Can you take us back to the start of this? Let's say that in this scenario that I'm about to ask a question about, you don't need any human input. From the time   that you walk the estimate with the customer and the time that you press the button on your pendant, can you talk us high level through what happens in your automations from the time it's recorded to the time that the estimate gets created in your estimating software?

Alex Boyd: Yeah, so basically how the automations operate and what's going on in each step. Okay.

Austin Gray: Yes. And the reason I'm asking is because in the video you said that you shared kind of the basic level, and then mentioned in the beginning of this call before we jumped on record, that there are some big time automation. 

‍Alex Boyd: There's an extreme level too. Yeah. Yeah. Let's go, extreme level through  the basic level real quick. The basic level is simple. The basic level is I record the conversation on the pen, like this little pen right here. it goes into plot. I run that against the prompt, then I'll copy and paste the sales notes from the prompt that we created into  the GPT we created. GPT that we created spits out the line items. then I'll copy and paste those line items into House Call Pro. And then the extreme level the part where you can take it even further is you stop the recording. You've connected the API to Google drive for your. plot, you stop the recording, it automatically uploads into Google Drive, and then you have an automation set up in Make or Na n. That will, in our  case, we use Make, it pulls the transcript from Drive, runs that transcript into a GPT, a GPT. That's the same. Prompt as the one implied, So it still creates all the notes and everything that we need. And then it'll then send the sales notes over to another GPT, which then creates the line item names, line item descriptions, and the pricing that then goes into parse. So it'll, I'll have that, I'll have the line items and descriptions output in JSON. So the GPT prompt is told to output it in JSON and then that JSON prompt is then parsed and sent over to a house call pro API a post API, which then will automatically post the estimate into that customer's   file, into their, essentially their folder file,which then automatically create the estimate for that customer. And the GPT kind of has error steps along the way where it says, Hey, you didn't include a customer name on here. so they say, oh yeah, so the customer's, John Smith. And then it says, okay, then it'll take that. And there's also a get customer note in there too, where it'll read the customer's name, get the customer's information so it gets the right link to post it on. But yeah. So essentially that whole automation runs by just turning off your recording. 

Austin Gray: Incredible. When are you putting a wrapper on this?

Alex Boyd: What do you mean what the Oh, to sell it, the automation. Probably never, it's my cup of tea. somebody else can do it. Fun. I like to do it, but, for me it's a hobby. It's, it just, and it benefits the business, right? That's why I get into the automations. I've got more automations too. I've got some other crazy ones for brushwork. 

Austin Gray: Yeah. What?

Alex Boyd: Yeah. So if you look at our YouTube channel and you see our shorts of the slow mo footage, when I, how I will post that is when I'm out on a job site, I'll record, I'll do a slow mo shot with my iPhone. And then I'll take that slowmo shop, and then I click on share, and then I click drive into my Google Drive folder. Okay. That's all I have to do to create social media posts across every platform is take the video, share it. No different than you taking a picture and sharing it to your friend. I have an automation set up where it's constantly watching for updates in this drive folder and then boom, new video comes in. That video then gets pulled sent into a, the link is pulled for the video, right? then that link is put into a spreadsheet. And then on top of that I've got a chat, GPT that's creating a tagline a title and a tagline for each one. Something that is consistent and works across every platform. And then on that spreadsheet, basically scheduled. There's another one that then a post to where it's reading that. Okay. It'll pull that title, the description, URL of that video. And then upload it into a software called blotato And blotato connects all of the different social media platforms.

So it'll be automatically posted to x, to TikTok, to YouTube shorts, to Facebook and to Instagram. And then I think I even have it going to LinkedIn. Because why not? And then on each one of those individual posts I tailor the posts for each platform. For example, Instagram wants hashtags, right? What I do is I give it a full, huge list of hashtags, then I have a JavaScript that every single time it runs, it'll randomly pull, nine of these hashtags. So the hashtags are never the same, but they're all relevant. And then 'cause that way it's just constant. It looks like, new, fresh content, like somebody created it and people do this already with AI videos where they'll create AI videos and then just load up social media with it. I was like why don't I just take this  concept and use real videos everybody's doing it with ai. So just use real videos from the job site. And I do the same thing with photo. No, I'm the one who posts the videos because I don't, the other guys don't even, I don't think they even know how to operate the slowmo mode on their iPhone. But I also have one for photographs. So anybody in the company that takes a cool photo on the job site, all they have to do is post the photo into a telegram chat and it essentially runs the exact same operation. And then schedules that post. So it creates a nice social media post just off of them. Oh, I like this picture. This one's cool. Post it in the telegram chat. Done.  It'll 

Austin Gray: I love it. 

‍Alex Boyd: 50 something day, however many days we've got however many days we're scheduled out on posts.  

‍Austin Gray: I love it. That's solving a big pain for people. 

‍Alex Boyd: Yeah. Like, why do I want to pay somebody to put up my social media posts? And make better versions by just training an AI to do it. You can take it so much further. You can have AI write every single, individual post, but you don't necessarily need to, especially  if you're just consistent content.  You can still post your one-off content. So if you've got something really cool or you want to your own spin to it or your own language to it, you can still do that. just that in order to keep your consistent content going, have an automation.  

‍Austin Gray: I love it. I absolutely love it. Yeah. What you can do with make.com chat, GPT automations right now is really cool and it's cool to live in this age. Now one thing I've seen that you're really good at. You're not only good at building the things, but you're good at training it to create good, consistent content. How do you train?   

Alex Boyd: Are you talking for like my personal X or are you like for the, like how do I train the content for the brushwork post?  

Austin Gray: Training is something I believe and myself included, I love building the things and seeing the automations get going, but training it to be better is the challenging piece. So how do you think about training?   

Alex Boyd: So I'll set up an automation at the bare bones level first, get it to work. And then go and look at okay, I don't like how this post looked on Instagram. Like this doesn't really match Instagram style. It doesn't really, doesn't have the hashtags that I want. It  doesn't I'll just literally just take  the output that I'm getting and then I'll send it into chat PT and say, I don't like this output. I want one that's more like this, and then it'll reformat the way that I have the JavaScript set up or something. And then it'll just. then I'll just essentially use chat GPT to train it, right?  

Austin Gray: How much time do you spend with Chat GPT each day?  

Alex Boyd: Not as much as you think. Yeah, not as much as you would think. If I'm, if it's a day that I'm sitting at the computer working I'll spend quite a bit of time fleshing out stuff. But really, I only use  chat GPT if I an idea  that I want to put out there. I don't use it to daily life or anything. It's strictly a tool for me for business. So if I'm working on something for business and I happen to be on the computer that day help it. I'll use it to work on it.  

Austin Gray: Nice. Nice. I love the content that you're putting out and I love what you're sharing with other operators. If you're listening to this podcast and you haven't met Alex yet, or you aren't following him yet. Go check out his YouTube channel. What's the tag for that? Alex? At Brushwork services.  

Alex Boyd:no at Brushwork Co.

Austin Gray: At brushwork Co. Okay. 

Alex Boyd: Yep. 

Austin Gray: He's  active on X as well. So if you're on X is it at B Print Co?  

Alex Boyd:Correct. Yep.   

‍Austin Gray: What was that from?    

Alex Boyd: So that's a Blueprint Studio Co. It was an idea that I had to, sell marketing and automations and stuff to other companies, to service industry companies. And I just decided I would rather just start my own service industry company. So I never really ended   📍 up pursuing that. But I kept the tag and I still utilize it sometimes like this winter I'm gonna be helping a tree service company, with all of their automations and their stuff. He's just a really nice guy and it seems really fun to do. So he'll get billed from Blueprint. But other than that, very few people actually interact with that company. 

Austin Gray: Now, is this a service that you're willing to do for other service business owners? Let's say for example, I were to come to you and be like, Hey, I want you to help me with creating an automation for my social media posts. Or, Hey Alex, I would love for you to help me build an estimating process. Is that something you're willing to do with Blueprint? I know you've never really mentioned it on X, I'm just curious. 

Alex Boyd: It'd be, it, is, but it'd be something I'd be doing more in the winter. So limited availability and limited amount of people that I would work with. So I'm, I typically, when it comes to the creative side of things it's a hobby for me. I love doing it. I love doing the videography. I love doing video editing. I love working in AI and creating automations. and I get really good at it. And so in order for me to enjoy it, I have to, if I'm gonna do it for other people, it has to be people that are open-minded and are willing to, listen  to or allow me to implement things that they think might seem crazy to them. But they've gotta be very open-minded. So I'm very picky about I'll actually work with. And I've always been that way. I've graphic design, web development, I've been very picky because it's a hobby and I want to enjoy it. I don't necessarily wanna make a living off of it. And yeah, so I would do it if it's the right people.

Austin Gray: Sure. Who is the if you do, if you could just identify the first ideal customer who comes to mind, what would that profile look like? What type of business, what are they doing revenue wise?   

Alex Boyd: That's the thing is to me it doesn't matter. It could be a big company, it could be a small company, it could be somebody just starting out. What's more important to me is that they allow freedom for me to be creative with it and do things that they think weren't possible. Or do things that they think seem like a crazy idea. Like the idea of automating your sales process from a little pen, from beginning to  end without ever taking a note or without ever looking at a computer screen. for a sales rep like that seems crazy to most people, right? They're just like, there no way. I can't do something like that. So the ideal person that I would work with are people that are open to those kinds of ideas and letting me implement them. So it doesn't matter. I've worked with big companies, I've worked with people just starting out. I've worked with mid-size companies. It just has to be a cool project. Something that seems fun to me to do.  

Austin Gray: Cool, cool. What platform are you most excited about? Like I've watched a bunch of your content and it seems like you have a really good pulse on what each algorithm likes. Which one are you most excited about?   

‍Alex Boyd: As far as social media platform or any platform like automation  platforms?

Austin Gray: As far as social media platforms.

Alex Boyd: That's a tough one because I feel like every social media platform has its own niche and, I think that X is really exciting.  I would like to but see that do better. But it attracts a certain audience. And the same is true for Instagram. It attracts a certain audience. YouTube is more, it encompasses everybody. but as far as social media platforms  go, I don't really have a favorite.   And I'm not really, I think that eventually somebody's gonna make a platform that filters out all of the AI stuff, all of the auto posting with ai. And I think that soon  in the next five, 10 years, you're gonna have a platform that is a human platform, and then every other platform is gonna have all of the AI generated stuff. And I think that'll be something we see.  

‍

Austin Gray: Specifically for brushwork customers. Like you've said, hey, you really enjoy creating the content, editing the videos, coming up with the idea posting about it. Which one are you most excited about that will help Brushwork grow will help the company. 

Alex Boyd: So I think it's gonna be YouTube ultimately. I think that Facebook is where we get most of our customers currently 'cause that's the best place to advertise locally. but I think that YouTube is the ultimate the ultimate one to do well with. And because if you get a big enough channel and people like watching you enough, like everybody knows who Matt's off recovery is, right? Pretty much everybody that watches YouTube knows who they are. And for a company to excel on YouTube, the sky's the limit. You can do whatever you want at that point. 

Austin Gray: So we were talking about this before we were on record. Yeah. I've been viewing our YouTube channel specifically for the Bear Claw YouTube channel, specifically for showcasing our projects. I'm trying to think about our ideal customer. What would they want to see from a company who they're going to work with? I thought it was interesting you posted this  automation video  on your Brushwork YouTube channel because for me, I've separated two, two YouTube channels. 'cause I have my own personal one, which is where I share my business related content because that's who listens to the podcast and that's who engages on X and I view that separately than what I'm doing with Bearclaw. And so I thought it was very interesting and I'm just curious to hear your thought process behind that and why you think that will actually benefit you by sharing some of that.

‍Alex Boyd: Because the end of the day, people don't buy what you do. They buy why you do it. when you're representing a company and you're, a company needs to have an identity, it needs to have a personality. So mixing your personal stuff, to an extent obviously but mixing those things together, you essentially create a following to the company, right? If you're trying to build a personal brand only and you don't, it's separate from your company then maybe I would, but people. Want to interact with your customers and want to interact and see a real person. They don't necessarily just wanna see a constant flow of job videos, right? They want to see that there's a person behind this and this person is intelligent or friendly or creative, or they want to connect with you as a person and your company is just the catalyst to, to make that connection and how they interact with you. 

Austin Gray:  Yes, a 100 agree with that. You shared something before, we're on record this earlier where, so this YouTube video that I'm talking about, and I'm recommending people go watch if you're a business owner, especially in local services space. Alex walked through the process of how he's recording all of his estimates and how he's using Chat GPT to or using AI to  create his estimates for him. But he posted that on his business channel. And something you said was, Yeah you just and correct me if I'm wrong, but you just think. You are letting your customers in to your process that you've built, which is your personality, and you're also showcasing the fact that you've built a pro professional estimating process, which I took that as your building more trust with your customers.

Alex Boyd: Yeah. And that's another thing too. You're building trust. Them an insight into who you are and how the company works. And they're. when you send out an estimate that you don't necessarily have it by per day something easy for them to  understand. they see that there is a  process and a reason behind it not just randomly throwing darts at a board to give 'em a price. There's a reason behind it. I think that's why I put it out there. And I'll continue to put those kind of videos out there. There's a I'm planning on going to the Paul Bunion show up in Ohio this year. It's in a couple weeks. And I'm planning to make a video of that. I've, our experience going to the Paul Bunion Show now it's relevant some way to brushwork but it's also very personal and the customer's gonna see that and say, wow, they're. They're going to these things. They're going to these events, they're keeping their pulse on the industry. And these are professionals. But it also allows them to see us and our personalities and what makes us tick,    

‍Austin Gray: Cool. Cool. Now to wrap this up, 'cause I know you've gotta get to the rest of your day, can you share your pricing or is that proprietary as far as tier one, tier two, tier three. Tier three and the combination of everything.

Alex Boyd: So what I will tell you is that there's a base price per quarter acre, and the tiers are multipliers. So are I won't share the exact pricing. but the tiers are multipliers. They're not necessarily a dollar amount increase. 

Austin Gray: Okay. Can you share what those multipliers are?

Alex Boyd: I would have to look exactly, 'cause I think it's tier three is like a 25% multiplier. Tier two is like a 10% multiplier. And that was, we found those multipliers by measuring the amount of hours spent  on certain tiers. The base rate is irrelevant anyways because it's gonna change for everybody. Your base rate where you're at, if you were to take the same pricing model, would be entirely different than mine. 'cause your brush is entirely different than mine. Your goal of clearing the brush is entirely different than mine. You might clear three acres in what we would do at one acre in the same time. And that it could be both of our tier ones. It's just different where you live.

Austin Gray: Absolutely. And for listeners, our equipment is different than what Alex operates. We do have a skid-steer with a mulcher, but the bulk of our work is done with a track chipper, feeding dead trees into or feeding dead trees with a mini excavator into the track chipper. And so for, from our perspective, what I've taken from this conversation is we need to track the amount of hours. That are allocated to each job and then document that job with before and after photos of density to train our GPT. Correct? 

Alex Boyd: Yeah. Yep. And then you can use terrain maps too for terrain, so you don't necessarily have to use just photos, but and once you get enough of that data, you can break it all apart into finding what's gonna be profitable no matter what. 

Austin Gray: Now, if there's somebody listening to this, and I'm asking this from my perspective, if we have not documented those exact hours on each job, but we do have the estimates and we do have the schedule. So we have a rough per day amount that we allocated with our crew and our equipment mix to that job. And we know the cost that it came in, and we've job costed all of our jobs. Is there a way to train CHAT GPT off of all of that data? 

‍Alex Boyd: Yeah, absolutely. Yeah.  So you can I mean you at least got a rough idea of how much time you spent at the job. You've know, if you're at a job for four days, you probably got it on your schedule that you were there for four days, and then you can just go back and if you don't have the area size, you can trace the area size say, I was at this job four days. Steepest grade is X, and then just kinda that together.

Austin Gray: What else would you share with listeners if they're going to create their own internal pricing model? 

Alex Boyd:Test it. Make sure you test it. Don't just do it. And you don't want to get pricing wrong. Stress test it for sure. Give it off the wall stuff, try to break it. 

Austin Gray: Perfect. Perfect. Alex, thanks again for being on for another episode. I do appreciate you sharing your advice.

Alex Boyd: Yep, no problem.

Austin Gray: And then where can people find you in brushwork? 

Alex Boyd: So you can find brushwork at brushwork co on YouTube. co.com is our website, and then me personally on X is B print Co.

Austin Gray: Fantastic. All right, listeners, thanks again for listening to another episode of the owner ops podcast where we talk all about growth strategies for growing local service-based businesses. We publish episodes every Friday and send a newsletter with a recap on Saturdays. If you haven't signed up for that yet, you can go to owner ops.com/newsletter. And then  finally, we've been documenting some of the free video   trainings. Inside of a skool group that will help early stage business owners get up and off the ground and grow their business. Th that is at skool.com/ownrops ops. That's skool.com/ownrops. With that, we're gonna wrap up this episode and we'll see you guys next week. Don't forget, work hard, do your best, never settle for less. 

‍

‍