Leveraging AI to Supercharge Your Company's Marketing

The better your business is, the more AI will do for you. And so by opting in and sharing your data with these models, because that's how it works, you feed more data in, it can do its job better.
If you're feeding in data, it'll do better. And really better businesses will thrive in that because you're going to close the phone calls.
You're going to have a higher ticket average, higher conversion rate, better rehash skills to recover some of that revenue. You're going to sell more service agreements.
You're going to use more financing, which we call promotions. Welcome to the Home Service Expert, where each week, Tommy chats with world-class entrepreneurs and experts in various fields, like marketing, sales, hiring, and leadership, to find out what's really behind their success in business.
Now, your host, the home service millionaire, Tommy Mello. All right, guys, welcome back to the home service expert.
I'm your host, Tommy Mello. Today is an excellent day because we're going to learn all about artificial intelligence and what it means to our business.
And most people aren't talking about this because they don't understand it and they don't know all the use cases. So I've got Jason Shandy and I got my man, Johnny from Service Stein here.
He works with the ad products at Marketing Pro. And Jason happens to be an expert in digital marketing, Google ads, SEO, data science.
He's out of Oklahoma. Johnny's out of Utah.
And we're going to talk a little bit today about how Jason is using AI for, he's the head of marketing at Southern Air, a $400 million HVAC company. And they partnered with Apex who does $1.3 billion.
And Jason's extensive background is in digital marketing. He's just mastered it.
They spend over $6 million a year in digital marketing. And he's used AI, ChatGPT, a lot of other things to build a company that actually thrives off of AI.
And I know a lot of people are talking about what is AI going to do? What is it going to change? It's actually very exciting. I think this is something that needs to be talked about.
So you're going to hear it here first. Home service expert talking about artificial intelligence, what it means to your business.
So first, why don't we just start out? We're going to do something fun here in a little bit. We're going to build a business.
Let's build a gutter business. Let's build some manuals and some advertising, how much we should spend budget-wise, some SOPs.
And we're going to do that for you guys to understand the power of AI and what this is going to mean to the industry. Because it's scary for a lot of people out there, but it's not scary for me because we're adopting it.
I actually have a full-time programmer working on things for AI, whether that's review response, whether it's rehash, whether it's understanding what our marketing bid should be. Should we be spending money on LSA, turning off our spend on PPC? Because PPC is the most expensive out of anything.
You've got four algorithms, right? Organic, GMB, which I love because we dominate those. Then you've got LSA, which is the next best.
Then the PPC, which is great, but it becomes expensive. One of my buddies, Lance Bachman, bought an HVAC company and just ran into correlation between humidity and weather into this PPC and figured out how to dominate.
He's like, it's so simple, no one's doing it. But when you could overlap data and understand how many calls you're getting based on humidity and the weather, some pretty cool things you could do in the HVAC industry, probably a lot of others.
So we're going to build a hypothetical gutter company, and we're going to show you guys exactly how we do it. It's going to be amazing.
First, Jason, let's talk about a little bit your background, who you are, what you're passionate about. Yeah, so I am the director of digital at Southern Air.
I got a shout out to Stephanie. She is the director of marketing.
Great branding. It's the year of the brand.
We've been saying that. We are heavily digital focused, but shout out to Southern Air Marketing team as a whole.
I do have an IT background, networking. And then I got interested in Google search about 2010 and got into web development from kind of back-end programming.
And yeah, I just got really fascinated by local search. And I hooked up with a guy in Oklahoma City named Ron Gregory about two years ago, 2021.
He owned a company there called Air Comfort Solutions. It's the largest HVAC company in state of Oklahoma.
They have an Oklahoma City and Tulsa location. So I took them on to do digital for them to take their marketing on, their digital marketing.
And they were acquired about two months later by Southern Air. And so I was fascinated by the scale.
And I really, I had never been exposed to the trades. You know, this is right after COVID where you could just see the resiliency in this industry.
And so I was offered a job at Southern Air to take on essentially all marketing channels as one of their regional marketing managers. I had 13 stores that I oversaw all marketing for.
But it was from the beginning with our CEO and our COO like, hey, I'm a digital person. I certainly will oversee these buys and we'll do this.
But it was from the very beginning so I can look at digital and kind of have a crack at fixing some problems that I saw were kind of endemic. And frankly, we had 46 locations.
And so when I came on, everything was outsourced. All of our management of our digital marketing was being managed by external agencies.
So I saw kind of an opportunity to develop a team at Southern Air and form the Southern Air Digital Marketing Agency, essentially. So we are a subset of the Southern Air marketing team.
So yeah, it's been a really cool journey. We started by taking on all of the paid ads.
So we took on Google ads specifically, like you've mentioned, because that was why. Because of the pain points in Google, pay-per-click is always typically the worst one.
your spending for these leads and these calls. And if the accounts are running unabated,

it can really do a lot of damage as far as cost goes. So yeah, we took a data science approach to it.
And that's what we've been doing for the last year or so. So how much approximately were you guys spending in PPC? So on Google Ads, we spend about $5 million a year.
Google Ads specifically PPC. LSA is around a million dollars a year.
All right. So Johnny, you are at Service Titan.

You guys are building all kinds of fascinating stuff with the ads.

And when people spend money and you guys have a huge data set.

I mean, there's not a lot of companies that could compare to Service Titan as far as the

money being flowed through that CRM.

Tell us a little bit about your background.

Yeah.

So as a college student, I took an internship as a marketing analyst. No idea what that meant.
I just wanted a job. Ended up being a PPC, Google Ads, for an electronics company selling HDMI cables.
So I started out doing Google Ads for e-commerce. And I'm kind of a data guy guy i really loved being able to see how much you're

spending exactly how much revenue profit that you're getting out of it and to me that was

marketing performance marketing was driving profit with ad spend then i got a job at an agency that specialized in lead generation like home services and i was completely blown away when i found out about this cost per lead model where you don't get to see the revenue. It's you spend this much per phone call.
And I would ask, well, what if it's a bad phone call? And they said, well, our job is just to make the phone ring. And it bothered me for long enough to where I worked with a team to solve it.
and we built an integration into Service Titan to connect ad spend with revenue.

That... bothered me for long enough to where I worked with a team to solve it.
And we built an integration

into Service Titan to connect ad spend with revenue. That was interesting to Service Titan.
They asked questions about it, and that led to a job at Service Titan on the Marketing Pro team to build something similar with them to connect revenue and data, to allow anyone spending money on PPC to see the whole customer journey, the actual penny to penny outcome of that

spend. And it's since turned into how do we apply data science and artificial intelligence to help our customers who have really struggled, especially with how expensive PPC is getting with click inflation and spam and just how competitive it is.
It's very hard to win at PPC, but we want to empower Service Titan users to use data and AI to do so. There's a lot there.
I mean, look, we've got 6,200 call tracking numbers in Service Titan. We've used CallRail and similar things things.
We use click cease to prevent fraud, but it still is by far the most expensive acquisition cost. I mean, it just is.
What happened is you look at a guy like Ken Goodrich, who decided to double down on branding, and he was spending somewhere in the lines of a quarter million dollars a month on ads. And what happened is he put all that money into branding, cut his PPC costs into a fifth.
And what he turned it is branded search terms versus non-branded. So instead of going for, I guess, air conditioning repair Phoenix or Las Vegas or California, wherever he's at, he's in a lot of markets now.
He started putting all the money into Gettle. And those keywords are a 10th of the price your quality score is through the roof and your conversion rate's higher, your booking rate's higher, and people are more patient and they don't shop because they're searching for your brand.
So I don't think branding will ever go away. Although I think you could win using data.
And here's the hard part for me is I've always wondered, I've got really good markets with great technicians. And if my guys are trained correctly, and that's the hardest, I think that's what brought value to A1 is we've got a training center.
We trained for two months. You guys saw we're watching films.
We're making sure our body posture is right. We're smiling.
AI will never be able to take, well, at least in the next decade. I don't think we're taking humans out of the field.
You know, what I want to jump into is let's just do this. We're going to build you a gutter company to the audience.
Now, there's a lot of people listening to this. So we're going to talk through what we're working on.
And I'm going to let these guys narrate it. And for podcast purposes, just so you know, we're building this.
So if you could watch the video, you could actually see the AI working and doing what we're asking for. But what we're going to do is we're going to build a gutter company.
We're going to come up with a business plan. Let's just, I'll mark down everything here.
We're going to come up with a business plan. We're going to find out what a good profit margin would be.
We're going to ask AI to find out our pricing model. So we're going to build out a business plan.
We're going to understand where we should

be at in the first year. We'll just do three good offers.
Let's make three offers that will really do well for a really good gutter company. We'll build a small manual for a CSR.
So we've got a marketing plan. We'll put three offers out there.
We'll find out what our financials should look like. We'll build a CSR manual.
We'll do a post to recruit a CSR.

You already said a general handbook, but various HR policies.

Various HSR policies.

Just positions like, let's define what is a CSR. Job descriptions we need to put on Indeed.

You know, we need the whole thing. So job descriptions for a CSR, for a gutter technician.
Yeah, let's do a gutter technician, maybe a dispatcher. And a dispatcher.
Okay. Probably need a general manager too, I would think, to run it.
Well, if you have Service Titan, you might not. So it's pretty interesting because the other day, my book came out right here.
If you guys don't know, it just came out March 15th. It's called Elevate.
And we were putting together the whole processes, the landing pages. There was a lot to put together.
And Jim said, this is two weeks worth of work. And he got on AI, logged on to ChatGBT.
And two hours, he put together two weeks of work. And it was solid.
It was good. And it's amazing what you can do if you're willing to take advantage and not be afraid to step outside of your comfort zone.
So we'll put together this plan. And if you guys get a chance, watch the video on how we built the gutter company because we're recording it.
We're screen shooting it. You're going to want to check out this video.
We'll send you guys the link. If you want to go to the podcast, just you'll see a clickable link here.
Let's dive back into this podcast. So with the rise of many AI tools in the market, I mean, there's a lot of them.
Google's kind of sitting by patiently waiting to come out with theirs, I think, of what I've heard. They're really waiting to see all the mistakes with ChatGBT because the first to the party usually is the one that kisses the frogs.
So I think there's going to be a lot of advancements in the next year, but it's a disruptor. What are some examples of the benefits that business owners enjoy when harnessing the power of artificial intelligence? I think, I mean, Service Titan is the perfect example.
If a company goes from pen and paper running technicians off of standard phone calls, off of a standard job board, and they implement Service Titan and use data uniformity, get on the data science path to where you can really start assessing how does your schedule look? You can really get a full 360 view of your business. And I think, could we discuss this? The first thing that these people need to do is get a CRM.
I think a lot of entrepreneurs overlook the value of something like Service Titan. And again, it's so complicated.
There's, like you said, there's so many different AI tools, but if you just rely on a big tech company, you know Service Titan, if they implement like some marketing pro feature, you know, like they are leading the way with AI. I was inspired into AI in the first place by Pantheon 2022.
I went out there and the whole theme was bringing AI to the trades. So I think the simple kind of roundabout way to get back to that is if you just partner with organizations or platforms

that you know are kind of leading the charge, you don't necessarily have to be a master of AI. You can say, hey, I know Service Titan and the engineers over there are doing cutting-edge things.
Maybe I can adopt them, you know, without breaking my business. You know, what always scared me is that I think a lot of people are afraid a big tech is accumulating all this data and then using it for other companies.
We're the largest garage door platform. I know there's some franchises getting on the service site, but I would never want my data and what service site learns being used for all my competition because I got a big data set.
What would you guys say? How is that kind of protected? At Service Night particularly, and I'm not on the data science team, but with the ads solutions that we're coming up with, we're very cognizant of that. We know that this is a concern that is shared by a lot of different businesses.
So very specifically, we are trying for the ads team, at least, to require some sort of opt-in to where anyone that opts in, you can share your data, see benchmarks, things like that. But we care about all of our customers.
We don't want to benefit one over another or favor one over another. You know, we don't take sides in the competition.
So we're trying to make it fair. At the same time, something important to consider with AI, and you see it a lot with Google Ads and its smart bidding, its own kind of inherent machine learning, is the better your business is, the more AI will do for you.
And so by opting in and sharing your data with these models, because that's how it works, you feed more data in, it can do its job better. If you're feeding in data, it'll do better.
And really better businesses will thrive in that because you're going to close the phone calls. You're going to have a higher ticket average, higher conversion rate, better rehash skills to recover some of that revenue.
You're going to sell more service agreements. You're going to use more financing, which we call promotions, but all the above that most businesses haven't even necessarily opted into.

You know, call tracking. There's so many businesses I see with one number.
There's so many businesses still out there, no offense guys, that are still using paper invoices, which I think is, it would be hard to manage at this point. It wouldn't even be a fair fight to go into that market with the data we have.
On every single technician and CSR and dispatcher, you know, we dispatch for profits. So we got a guy, half an hour, another guy, 45 minutes.
And there's a guy that's better at the house with three doors. We're going to send the guy that's 45 minutes away.
So it's not always what's best for every business. Because one guy might say, listen, we've already got set pricing for this job.
We're going to book the closest guy. We're optimizing for routes.
And maybe that's what you do for pest control and pull services. And there's other industries in that category.
But for us, we're like, listen, this guy's equipped to sell three doors. He's got a better foundation and the knowledge of a wood overlay door, what that might be.
So understanding what technician for the dispatch, what CSR. we do a round robin, but it's a weighted round robin.
So our top CSRs get more volume. They can make more money because wouldn't you want that at your business? Wouldn't you want the star quarterback to play in the game? Or do you want the third string guy coming in all the time? Hopefully not, unless you're Green Bay.
No, I'm kidding. Right tech, right call.
I mean, that's the foundation of everything that we do at Southern Air, essentially.

And you don't know what you don't know.

So when you come in and you implement these tools,

then you start seeing that bird's eye view

of things that you're missing before

if you didn't have these data points.

You mentioned in the book,

I mean, there are SMS drip sequences.

You know, that is an AI use case

that's been in use for a long time where you just collect customer information off of a drip. The account reconciliation tools, just all of these things that human beings have been doing for a long time that are tedious.
I mean, mass adoption of these AI platforms is coming to where the mundane little tasks that people are doing right now, the day is coming soon that it'll just be a prompt that is accomplishing it. You know, the question I have, and I'm going a little off tangent here because I'm obsessed with search engine optimization, and you asked me for a link earlier so you understand the power of link building and on-site optimization, blogging, coming up with the right blogs, the right keyword density and all that stuff.
What do you think is going to happen to organic search as we know it? Where are we headed with that? Because right now, I was just going to bring up Jasper. It was originally Jarvis and it changed to Jasper.
Yeah, so Jarvis. So what happened is they got sued and they had to change it to Jasper.

So I don't know, four months into it,

Jasper came along from Jarvis.

But what is going to happen with organic search

and content building and blogging?

Because I could write a book with AI.

I wrote my book with pure intellect,

which was probably stupid.

We were talking about it at lunch.

And frankly, I think the future of search in trades specifically is if it's Google or Bing, whoever it is. Like right now you have local services connecting.
Google connects the customer to the vendor. I think eventually you're going to have the software partner with companies like Service Titan that have thousands, a network of trade companies, and then that are feeding the right signals through the algorithms.
So having all of the data structured to where if I Google, hey, assistant, I need air conditioning, my air conditioner's broke, to where right now we're discerning through the SERP, but algorithms, AI algorithms, if you tell it, pick an air conditioning company. So I think it's headed to where you'll define a profile for yourself.
I am this. And then the companies that have their data structured in the right ways.
If I have a technician sitting on 119th and Western in Oklahoma City, and you live right there and you Google AC repair near me, and my company is tied in with Titan and Google and all of the lead aggregators, it won't even go to a dispatcher at my company. It will, service Titan alert from Google, go to this job.
That is very theoretical. Well, the issue I have with that is, it sounds great in hypothetical, is I always thought about the same system, having a group, a network of garage store guys.
And Amazon has thought of this three different times, and they failed three different times with Amazon Home Services. I ordered a chair two months ago.
The guy didn't come for two weeks. Then he was short a screw two weeks later.
It just doesn't work. So this is why, with your theory, I'm having dual cameras installed in every single truck I have.
Every guy's background track, drug tested. They got the right tools.
They got a stock truck. They've got the connections to service.
All of our guys know how to use the mobile. So I'm building an army, and I think tech will be just dependent on me as I am of them.
Because where else is there an army that has full coverage across the United States that can make 100 technicians a month? So the training center, how we use AI to train. And they can do voice sentiments.
They can tell you based on what I've heard over time, this guy needs to work. His eye contact wasn't correct.
There's certain sentiments. You could basically do personality profiling through AI.
And so the issue I have is if you took a bunch of companies, HVAC, and they're all using different price books, and they're all using different ways of scoring the calls. Because some guys say, you know, some people have different dispositions.
So by having everything uniform and getting more garage door companies on our price book, not the same pricing, but the same chart of accounts, all these things the same across the United States, potentially, and they got to have background checks and they got to have the vehicle cameras and there's all kinds of little things. We even tried wearing a camera for the tech.
Now the customers didn't love that. So we got to figure out how to do that.
But I think there's a real opportunity. I think you're going in the right direction.
I think potentially there's going to be sensors on HVAC, on all of plumbing, on all the garage doors that will know that there's a problem before the clients do. That's where I think the world's going.
The manufacturers are going to do it. They're going to pick the best clients.
They're going to go rep split. They're going to say, I want 20%.
I'll feed you unlimited leads because you're the only one that can fulfill it in that area. We know you're capitalizing it.
We know you got the highest

reviews and it probably wouldn't be reviews. It would probably be a sentiment that goes off of

net promoter score, right? It would be something more internalized,

but that's a deep conversation. It could go a million different ways, but

let's talk about how AI is disrupting marketing as we know it. Because, you know, with the Zapier that just came out with ChatGPT, there's personalized marketing.
It's got a bigger data set to be able to collect what is working. Yeah, and that's essentially what we're doing, is we provide personalized customer journeys.
So we looked at historical data, millions of dollars in ad spend, and essentially standardized best practices that we saw lead to conversions over time and systemized it with. So our landing pages are served post-click.
So we have a system in which when you click our ad, we pull some things from the click

identifier.

It's a technical thing.

And then we adjust our landing page post-click.

So if you...

It says your name, a picture of the Facebook prop.

All right.

It's along those lines, right?

How is Finnegan doing?

Indeed.

Yes.

So it is very perfect.

If you're looking for air conditioning repair, we'll show you $50 off.

And it will... The city that you're searching from.
If you're looking for a tune-up, we'll show you the $69 maintenance tune-up. So we have, we switch offers, we switch headlines.
And so we, initially it was, I got to build 50 sets of landing pages to run campaigns. Well, to do that, then I got to build 100 landing pages per brand to have that granularity that best practices require.
So that's where we got into developing the AI system that says, well, what if we just had one endpoint for air conditioning for all brands that you feed signals? So somebody is looking for Goodman HVAC or Train or Linux. Is it programmatically showing that product for if you sell that product? So yeah, well, I mean, we certainly can.
We don't actually bid particularly on branded, those branded keywords. They don't convert super well.
But a better example is like AC installers near me and the hero image will be picked from a series of images that are essentially competing with each other. And it will be like train images.
It'll be various images that we have fed it. And so we have a system of ranking that is constant.
So whereas before you had A-B testing, we call it X testing, that every possible combination of landing pages that could ever be produced is produced once that signal is missed, and then we rank them in cash. So that was a very, very poor way of describing it.
But basically, we are constantly assessing what is the best air conditioning repair landing page at all times in all markets. And post-click, when a click happens, there's a programmatic check done that says, what

is the best headline for air conditioning repair?

What's the best offer?

And we are feeding it.

You mentioned weather signals.

We have all sorts of signals going into the algo that is producing landing pages live

at the time of click versus the traditional way of doing it, which is guessing, you know, which ad groups, which keywords, which landing page. I get that.
So if they're on desktop, it's probably a little harder, but you have a click to call that the AI recognizes that that button was hit and you could actually manage the fact that this actually did turn into a phone call. Form leads will be the easiest way because you know they left the lead.
So we track it on desktop and mobile. So on desktop, we are just going to show you the schedule engine integration.
We show you the phone number. But yeah, I mean, if you manually dial the phone number, we are using the Service Titan DNI.
So it will capture the phone call and then the click is captured in Google Analytics. But yes, you can track that interaction.
Yeah, because you want to have an end-to-end feed, like a loop. So can you share with us a little bit about Southern Air and how it's using AI to perform predictive analysis and campaign management? We went into that a little bit.
I understand that, but there's probably more to it. By analyzing millions of dollars of historical ad spend, you're able to engineer a system that just is probably better than most people's.
I mean, as you go into that, what's your target percentage of revenue to spend on marketing for PPC? So we do it. We don't do a percentage to spend.
We do it just return on ad spend. And we try to be eight to 10X.
Typically the case study that we ran with Service Titan during shoulder season, we've maintained 12x. So, a 12x return on that.
That's incredible. So, if you were at 10%, that would be a 10x.
12% is about a little under 9x or 11x. That's why I had to bring this guy with me.
And honestly, the truth is, when I built it, it wasn't to make the best possible landing page as far as performance goes. It was honestly, how can I do this myself? I have to manage 50 accounts all at once.
I can't make 10,000 landing pages. What is the most efficient way possible? Should I build 10,000 landing pages? And it's just, I guess that's the, how kind of innovation happened in my own head to say, you know, there's a better way of doing it, not just to increase performance.
That was an ancillary result. It was to, you know, how can I manage it with, in the most efficient way possible without a team of people and creative and web developers? To get a 10X on PBC is very hard to find people that are doing that because I use it.
I've always used PBC for capacity, right? I need enough leads to keep my guys busy, which is three to four leads a day. PBC is the last thing I want to spend money on because it's the most expensive.
I mean, it's more expensive than a ValPAC for us, which is pretty crazy. So you guys are doing campaign performance.
You guys are doing LSA. How are you using it on LSA? Well, so you kind of touched on it earlier for budget trending.
So if we know we want our LSA profiles spending just wide open at all times. So if you set it at $10,000 a weekly budget, then it will spend $10,000 if capacity is there.
So we have to be careful, especially during peak when our boards are full. If our LSA spend goes up, there needs to be a inverse drop in our Google Ads spend.
So we have a machine learning algorithm. So that's all doing that behind the scenes.
Yes. And we adjust, we pull levers on the Google Ads side.
As LSA spend goes up, There's a finite bucket of total Google Ads spend made up of local services and Google Ads. And that's what we use our algorithm for is as Google LSA spends more, PPC must spend less.
So step one was just turn off ad groups, turn off lowest performing ad group at the time of signal. So then we got smarter, right? Using various

algorithm methods, we got smarter to say, you know, well, only turn off this ad group if this

is happening. So we use conditional logic to assess real time.
And the way that I described

it to our CEO, if you, my CEO called me and said, hey, our phone is ringing off the hook right now.

Local services is spending crazy. We need to turn Google ads off.
I would go in, check local services. Okay, I have $10,000 left to spend on Google ads.
Where do I spend it? I would do a real-time analysis. And so that process in which I would do my real-time analysis, I've done with the data science team for the last 18 months.
Everything that I have done in these ads accounts, everything has been systemized. We have sheets, we have changed histories to where I have a data science engineer.
Every time I do something, I explain it to them. So we have been synthesizing my process.
So as we have been trying to pull levers, LSA, PPC, how do we keep this balance going? It's just come about like that to where, you know, I guess, what do they say? Necessity is the motherhood of invention. I'm thinking there's probably something, like I said, potentially I see this thing saying, if you built these five landing pages with your domain ranking, you would be able to get this many more phone calls based on search file.
And the problem is Google doesn't tell us anymore necessarily what people are searching and how much they're searching, not like they used to. So the Google Council used to be able to see there's so many searches for this.
You could have the content built with keyword density and rank organically. I'm still wondering what Google is going to do about organic search and with the content.
Because is there any way for them to tell my buddy that really is the godfather of SEO, Matthew Woodward said, they'll be able to detect AI. And when they do, you'll get penalized.
And if you get penalized, there was something that happened years ago with JCPenney. Before the old way was spinning content.
You'd write the 10 sentences, put in every synonym in there, it would change the words out and you could spin up pages like crazy. That was really easy to detect.
Any algorithm these days could pick that up at 10 years ago. So now the question is, will Jarvis and ChatGPT be able to get picked up by the algorithm to say this isn't unique developed content? But I don't know.
Is that something that you guys think is going to happen? So Google's recently come out and they've said, we're not going to penalize. Pretty much they said we can identify it because people have been cheating on tests, right? They've been writing essays through ChatGPT.
And I don't remember the name, but there's a software company that grew out of that that can detect if it's ChatGPT generated content, give it to professors and professors run it through to make sure it's not generated. So Google can tell, they've said we're not going to penalize it.
They said we have kind of our checklist of what is good content. And if you did some of it by AI, that's great.
That's what they say. It's a bit of a pivot from where they were a few years ago.
They're saying, if it's helpful to users, then we will not penalize you. It's essentially a doubling down of before where they did say the helpful content update that Google uses AI for everything.
So they made it clear in there in the policy kind of outline that they released a few weeks ago that if it is legitimately not trying to manipulate search, we just did an article and I used AI to do it.

It was, I need the anatomy of an HVAC system.

So we used AI for the outline and then we rewrote the article.

But the illustration is, I don't know.

I mean, I do run digital at a very large HVAC company and I I know nothing about the anatomy of an HVAC system, and neither does my marketing team. But I told him, hey, go write this article.
He used AI to write it, paraphrase a little bit. But even I can say unequivocally, they can detect AI written content.
But they've made it pretty clear that it just wouldn't be fair to punish people for using it because it is everybody's using it. You know, I look at 10 years ago, we had a for I would you have a for my master's program and I had access to a plagiarism checker.
So I'd run everything through because half the shit was a bibliography of what I, you know, as long as I cited everything, it could be half. As long as I cited it, we're excited on the back.
It was okay. But I'd be like, man, that's a red flag.
But you could read everything and reinterpret it in your own way. So I think it's a wave of the future.
I'm actually 70% excited, 30% worried because now it'll start the advancements. I mean, you heard Microsoft come out, they're putting another 10 billion behind it.
And if that's the case, it'll start developing code. And it'll get smarter code.
It'll do like Ruby on Rails and Python and crazy stuff. And it'll start building apps.
And it could probably duplicate. I don't think you could ever duplicate a service titan.
But you could easily build a dating app that would be better and take more interest of the person and take all the stuff online to build. I don't know what's going to happen.
It's pretty crazy. What made you look into using AI and you said you've been doing it about 18 months? Pantheon.
Pantheon. That really was.
So I went, we went to the service site in Pantheon and I saw Johnny speak.

And I forget what you were talking about, the Google ads, the marketing pro Google ads integration. And I've never met anybody that was like as obsessed with Google ads and HVAC than me.
And by the way, he's tall. So I'm six foot eight and he's like six foot five.
So we are kind of a funny site of Google ads giants, for lack of a better term. But yeah, so we got acquainted there.
And I love the notion, bringing AI to the trades, just the thought that we could put software automation tools in place. And it sounds cliche, but really make a difference in contractors.
And I do represent PE and it's this weird dynamic that I, you know, I am big corporate America, but I think my real fascination with all of these automation tools is I know that it is going to empower these contractors to stop getting taken advantage of by tech guys. Because the truth is, digital marketing people, we have a bad reputation in the space because people come in and they overpronged, we're going to, your phone will never stop ringing.
We're going to build you the coolest website ever. And it's just the nature of these relationships that we've always me and him kind of both believe you should do your own marketing if you can.
You know, and these type of tools make it possible before you had no chance of building a good website that has content that is structured, like you said, for SEO, but you can spend the time now. If you really are dedicated and committed, you can do a passable job on your own digital marketing using these software tools.
And it's becoming easier and easier. And that's my honest passion with it is make it to where I can call a plumber that he just left his job at such and such, and he's going out on his own, and he downloads Service Titan and whatever else, TechStack, and he's running Google Ads.
He doesn't have to call a data science expert in all this technical stuff to make his phone ring. I think that we're living in an age with uncertainty with what's going to happen in the economy.
But I think the problem that I see is a lot of companies right now, they're calling their marketing agencies saying, I need more leads instead of fixing the root of the problem. You're not booking enough phone calls.
You're not selling enough service agreements. Your ticket average is not the real cost.
If you analyze all your bills, the air conditioning bill, every bill you pay for, for your tablets, for service time, for your vehicle payments for for all the insurances, for your lawyers, for every CSR dispatcher, you add it all up. Most of the time you'll realize you need to charge 500 bucks an hour because you're not working 40 hours in a garage.
You got drive time, you got your meetings, you got your dispatch time. You've got all these different things.
You're actually working about 18 out of 40 if you're lucky. So if you're paying your guys a decent wage, I mean, if they're making $1,500 a week, and that's usually performance pay, then you got to add in the parts costs.
Then you got to add it in their time and your CSR's time and your warehouse guy time and your accountant's time and your marketing team's time and the cleaning lady's time. And it all adds up.
And there's no formula I've ever seen that got less than $400 an hour. A lot of guys are at $650 an hour.
And I don't think people understand if you want a warranty, if you want a drug tested background check guy to show up to your house and do everything for you, diagnose it, fix it, warranty it, show up in and out to save around your kids. That's how much you got to charge, especially if you want to have meetings and help educate your people.
And you want your people to be homeowners one day and you want them to be able to go on vacations. And I think a lot of businesses, unfortunately, and I'm not talking smack because I was this guy, but a lot of guys are running a business with two employees and they are the business.
So they could afford to do things that a lot less. But when they go on vacation, their business falls apart.
They can't leave their business. It doesn't run without them.
And unfortunately, they could charge less, but it's not giving the homeowner the best experience when they call up in three years and you're out of business or you're on vacation or you might have cheapened up on the parts you're doing to give a better deal. So when you start becoming a big business, I think you start delivering a new expectation of service and it doesn't run with the owner.
It doesn't need to. It works without them.
But I think AI is going to make this a little bit easier for small businesses. The problem is big businesses are going to go a lot deeper in the paint the next year on this.
And the small guy has got to figure out a way to keep up. You mentioned in the book, the story, Alan, this guy, Alan, that he's doing $6 million a year, and he ends up selling out to PE for, or maybe it was $5 million here, he sells out for $6 million.
PE comes in, they adopt systems, they grow it to $22 million the next year. These tech stacks will enable that guy, Alan, to do this himself.
What we are doing at scale at PE, we are taking advantage of the software tools in the most efficient way possible. But as they get smarter and easier to adopt, Alan will be able to do it himself.
I mean, I truly believe that. You could make a cool website.
You could literally tell your AI assistant, I need a website for my HVAC company. And that is how it will be built.
That literally is where we are now. I mean, it feels like kind of a distant future, but I mean, do you agree? Yeah.
I feel like you could probably hear some of the stuff Jason was saying earlier about the really cool applications that Southern Air is doing and thinking maybe you're like a $2 million shop thinking, I can't hire Jason. I can't build a data science team.
I can barely keep up with what's going on. So how do I apply AI at my size? I think of when smartphones came out, you kind of had to pick, am I going to be an Apple person? Am I going to be a Samsung person? Heaven forbid, a BlackBerry person and shun it.
You almost like hit your wagon to a company, a brand, and trust them to get you the best experience possible. Realistically, most people listening to this are not going to hire an AI team.
You're going to find a tech stack, to what you said, Jason. I mean, everything you talked about with the budget sharing workflow, just for example, obviously I'm a Service Titan employee, so go Service Titan.
We are on the ads team building an almost identical thing, a recommendation engine of here's where you should shift your budget to the most profitable campaigns. It's on the roadmap for our new Optimizer product that's coming out.
What about when? That one, recommendations for Optimizer sometime this year. You know, I called Vahe, you know, the founder of Service Titan, Himanara.
And I talked to him, I don't know, last Wednesday, Thursday. I was really calling him about this whole Silicon Bank thing that just big tech had to deal with, but Service Titan was smart.

But I said, man, I need a lot more access to the V2 API.

There's certain things that I can't do that I need to program to be able to run my business.

And he said, well, I got a hold of animal.

And actually, I owe him a phone call later.

But no, you're right.

The hard part about technology for me is we've got a lot of people that are using Microsoft Teams. We've got a lot of people using Zoom.
We've got a lot of people using Microsoft Word, sending docs via email. I could go on and on.
Some people love the Trello board. Some people are using Monday boards.
Some people are using Asana. And so I think the main goal of every company should be, here's our five stacks.
We actually use HubSpot for certain things. And there's this book called Come Up for Air.
My buddy, Nick wrote it. And it's all about just getting your tech stack right and getting everybody affluent on it because he could guarantee you just through email alone to save everybody one day a week.
Everybody in your company. Our CTO was looking for an NDA.
It took him two hours. If we just knew what technology stacks, searched NDA, pulled it up, the one that we use.
Same thing with manuals, like get them all in one spot, get every piece of data. I don't believe in email internally.
I think email is a big waste of time. I think it should be project-based in the Sonner Monday.
And you assign tasks and we know where everything's at and it's all we could. Just what I loved about Service Day the first day I got on it is I knew everybody that touched it.
You couldn't delete anything. I knew this dispatcher touched it, this technician came and touched it, this installer.
It's a data set that you can't spam. So went off tangent a little there, but I think there's a lot of businesses that need to decide on the technology, train on what they're going to use and make sure they communicate properly.
AI could be used for the greater good, but it could also go the opposite way. And this is the scariest thing for anybody thinking, man, this could be used for evil.
What are some of the scary things that you guys see AI doing in the future and that you're like, pray to God, it doesn't happen? All brand bias, especially in this space, is what are the systems that are going to select what companies rank? Just like right now, the Google's dominance of the cert page, what is it going to be like when it's reduced even further from seeing 20 different selections to you just have a verbal return of call this company? Then obviously it's the work-related, but then not trade-related. What if I'm trying to learn, you mentioned, what if I want to learn Ruby on Rails? I'd go through an AI guided course and it teaches me you know there are biases that it can introduce that you know with politically so that on and that's what I was gonna say that's for like the AI ethicists greater minds than mine but I do certainly worry about what it's gonna do to search what about you yeah I'm gonna go a bit more existential you know like robots are going to take over one day and dominate us but i'm worried about the bias because chat gpt what it did like why it exists is it it kind of learned the way humans do right like there's different patterns that our parents teach us and we start to understand the english or whatever language we know.
And then we start to form more complex thinking. And that's what ChatGPT did, but with almost infinite brain power so that it could just learn everything.
But it's based off what's already been written, based off of previous human ideas, which same as we learn. We know because someone taught it to taught it to us but we are biased imperfect people and we've taught chat gpt how to think in a way so chat gpt has opinions about political issues um has they kind of had a political when you asked about global warming it said all the reasons it doesn't exist and within two days days, it changed its tune completely.
Like somebody overread the code. That's the concern, right? Right.
That's the problem. Google has been doing this for years too.
I mean, it's spoken of less, but I think it's known in search engine marketing circles that, especially with the political stuff, it's very slight bias where

you would look for articles and they would serve just very slightly biased versions of the news

story you're looking for to where it is still Google search and you're still seeing 10 articles,

but the number one appearing article, which they know is the one your eye draws to, is some sort

of slanted. So again, it's all conjecture, but these things have been happening for years anyway.
So the concern with all these tech companies having all the power that they have. That's exactly why Elon Musk got involved with Twitter and their Twitter files coming out.
I heard Facebook came out with an AI tool and it began to kind of run its mind of its own. They had to unplug it.
I don't know how much truth there is. You know that story? The AI developed its own language to communicate with itself.
And at that point, they had to shut it down because they're like, this is end of the world type stuff. So that's what freaks me out is like, we don't know what we don't know.
So now that we have these algorithms that are self-improving and as computing power just keeps getting more and more powerful, we don't know where it's going. Because like you said, they will start building their own software applications.
I watched an interview by Elon Musk the other day where they were asked, you have several hundreds of billions of dollars. What keeps you ticking? Why are you at work still working 18-hour days? And he said, I've questioned the meaning of life.
And the meaning of life is actually out there, outside of just our Earth, outside of our solar system. He says, so we've got to go far beyond just Earth.
We've got to have the human species outside of our solar system. And he says, I think we could get there.
He goes, that's why I say we need more human beings. That's why we need more population, which is the opposite of Bill Gates.
He says, we need human beings to collapse. The ecosystem and our planet will not support human beings.
And Elon just completely disagrees. What are some of the challenges of setting up AI tools? I mean, as far as the average Joe, what are some of the obstacles and where do you think it becomes easier? Obviously, you can call service day and get a hold of Johnny.
But what's your next steps? What's your next three moves? Selecting the right person to help you. Certainly, if you're not a technical expert, where it stands now, probably don't try to do it by yourself.
Make sure you find, Service Titan has like a list of certified providers. And even that, the AI section of that list is probably pretty small, but I would say where it is now, outside of using chat GPT and stuff for, you know, writing your manuals, kind of how we discussed and how we're going to do here in just a second.
But trying to get programmatic bidding, smart campaign optimization, doing all of those things. I would say wait until guys like us finish building those tools for mass adoption.
And in the meantime, definitely find a good partner to help you. And I'll say probably reach out to Johnny and his team.
They have good agencies that will teach you, hey, you don't necessarily have to pay us for 100% of your marketing. You can use these tools and that tool and kind of run part of it yourself, but essentially partner with the right people.
So there's two aspects of what you talked about with ServiceTight. There's the data science, regression testing, statistical analysis, and there's the AI.
What is that balance? How does that work? Because you've got the data guys putting stuff into a BI tool, trying to figure out different opportunities. How much is AI and how much, because you said it's a little bit of both.
Right. So yeah, well, the Marketing Pro Ads team, which is him essentially, and we conducted our beta test and then they had their data science team go through our numbers and assess performance essentially.
So that's what the report that they produced was their data science team. But we worked in conjunction with their marketing pro team that has the ads optimizer product.
And what it does fundamentally is feeds revenue data from Service Titan to our Google Ads account, amongst other things. But that's one of the big benefits of it.
And then that helps us in our bidding. You know, we know certain jobs, certain conditions lead to better clicks and better outcomes.
So that data exchange is very, very useful. How many large companies are on that ad optimizer, like PE led? Is there a huge data? Is there a lot of companies that have opted in? Yeah.
So this has been a closed beta right now. We have probably about 20 different shops in that initial.
What are the categories? Let me see HVAC Plumbing, Electrical. Yeah.
HVAC Plumbing, Electrical, one garage door, I believe. Okay.
Well, I need to talk to you about that. So listen, guys, there's been a lot of knowledge dropped here.
I think the audience is going to love the activity that we did on the video that kind of shows the process of setting up a company from soup to nuts. Obviously, it's a lot more antiquated than that, but this is just a little brief.
You could spend an hour, build a business plan, submit it to the bank, get a SBA loan and be in business. And for's for those out there, don't start a home service company based on just chat GPT.
But, you know, it's a lot of hard work. There's leadership involved.
There's culture. There's a lot of things that chat GPT.
Maybe it'll know when to send my guy a gift and what kind of gift they want. Or maybe it'll be a version of me.
What is that called? Deep fake? Maybe it'll do a deep fake version. Great.
Cha-cha-cha-cha. So, if somebody wants to get a hold of you guys, Jason, what do we do to get a hold of you? Jason.shandy at southernairnow.com.
LinkedIn, Jason Shandy. Jason Shandy.
It's a good beer. Ah, you know it's a lemonade beer.
Okay, my kind of guy. He knows.
And Johnny, if someone wants to reach out to you, what's the best way? Oh, I'm hesitant to give it out because I get all sorts of service titan tech. You don't get a hold of Johnny.
You got to get a hold of me or Jason and get a hold of Johnny. Let me ask you guys this.
If there's a book out there that changed your life, could be technology-based. It could be as simple as, I'm guessing, Men Are From Mars and Women Love the Venus for you, Jason.
But no, I'm kidding. But if there is a book that you guys would recommend to

the listeners, do you guys have a great book? Oh, my gosh. Well, it's not germane, but I was just

telling them about it. It's called What Every Body is Saying.
I read, well, I got to say Elevate.

I did read Elevate to the audience. Read it, right? But Jonah, what I was just telling about

What Every Body is Saying, it's like a study of body language, nonverbal communication. Because, you know, look, AI is going to study your body language.
So you need to be prepared for the robot war to, you know, be able to not give off the wrong signals. Yeah, there's a good book.
I was going to meet the guy in Vegas, Never Split the Difference. And this guy, Chris Voss, worked for the FBI.
And actually that's all he did is recognized. You look to the left, you look to the right.
It follows eye patterns. The biggest thing.
Now when I go through the airport, I just go through the, I'm pre-checked, but what's the other one called? Clear. So I just run my eyeballs through and boom, it's crazy.
I'm a little bit worried. I'm not going to lie.
Johnny, what is a book that you really like? It's a classic, but seven habits, highly effective people. Yep.
Listen, here's what I like to do. Obviously you guys know your stuff.
We talked about a lot of stuff. I want to give you guys an opportunity to close us out.
I'll give you each couple of minutes. Maybe we talked about it.
Maybe we didn't, but I want you guys to close us out with something on your mind. Maybe it's something that you need to go do.
Maybe it's just a last final thought, but you go first, Johnny, and I'll let Johnny close us out. Jason, Johnny.
So, man, I got to say, I did read the book and, you know, I think I had an aversion to the trades initially to your business model before I even knew about it.

The Air Comfort Solutions business model is very similar.

And that's a Southern Air brand.

My introduction to the trades, it was premium.

You're going to pay more.

Our technicians look different.

They are different.

They have better benefits.

Some of them make six figures.

That whole model, it was fascinating to me when I read your book because I think Ira Pruitt must have read that 25 years ago. That is the Southern Air message is we are premium all the way through.
I'm inspired by the collaboration. And I just am humbled and grateful to be here as a representative of Southern Air.
We haven't said much about them, but what an organization, hundreds of employees, my marketing team. It is not just me, to be clear.
Stephanie, Colton, and Tim, and just Ann and Jen. And don't get mad at me if I forget you.
But I'll drop from there. But thank you again for having us.
Yeah, this is great. Start the Oscar music, right? You can't thank everyone.
So I would just advise for you, like, this is cool. What do I take from it? Go Google how to get a chat GPT account.
Just start with that. Look it up.
Chat.openai.com. There you go.
It's free too. Get it.
Yeah, get it. Keep going.
Get it and then just mess around with it. Watch YouTube videos on it.
Dip your toes in that world and then let your creative business brain go to work for solutions. If you haven't heard about ChatGPT and you are now hearing about it, anyone that's listening, so remember the first time you ever saw an iPhone? That's that transformative.
So if you're not using this to develop systems for your business, to develop job descriptions and best practices, then you are already behind because we have been doing it for months now. So I'm going to close this out.
I asked Chad GPT, what's the best home service industry to be in? And she said, I named her a she. There is no single best home service industry that suits everyone as it largely depends on your individual interests, skills, and resources.
However, there are some ideas for home service-based industries, so I didn't really read the question right. But one is e-commerce with the rise of online shopping.
You can set up your own e-commerce store selling products to create or source from other suppliers. Two is freelancing.
Depending on your skills, you could offer your services as a freelancer. Third is online courses.
Fourth was blogging. Five was virtual assistants.
Six was catering. Seven was pet grooming.
Remember that starting any home service industry requires research, planning, and hard work. It's important to identify your strengths, target audience, and unique selling points to create a successful business.
Great advice. I'm Tommy Mello, the home service expert.
Thanks for listening. states.
The insights in this book are powerful and can be applied to any business or organization.

It's a real game changer for anyone looking to build and develop a high-performing team like over here at A1 Garage Door Service. So if you want to learn the secrets to help me transfer my

team from stealing the toilet paper to a group of 700 plus employees rowing in the same direction,

head over to elevateandwin.com forward slash podcast and grab a copy of the book.

Thanks again for listening and we'll catch up with you next time on the podcast.