AI Agents for HVAC Businesses: What Actually Works

Who is Jim Christian, and what does Podium do in home services?

Jim Christian: You're not going to lose your job to AI. You're going to lose your job to somebody else with AI. There are plenty of AI companies thinking about how to do full human replacement, and what I push on our team is let's focus on turning people at these companies into superhumans with AI, not focus on replacing them. So we built a full operating system platform that's not just the front office communication layer anymore. It's a full field service management software under the hood: price book, membership programs, the technician mobile app, the entire stack. The third one that is honestly the most exciting to me, we call it the Podium Operator. Imagine you had another employee or assistant in your office that could do work for you when you need them to. He has skills, a very human-readable playbook of how I should operate in these scenarios.

Tersh Blissett: I am excited to have you on the show today, Jim. Tell me who you are, what's your role. He's a friend of mine that I've had for so many years now, and literally I could text him anything and he's like, "Dude, you got to check this out, we're doing this new awesome thing." Back whenever AI voice was coming out a while back, I saw him at a trade show and he's like, "Dude, you got to check this out, this is awesome." Podium's been in the forefront of the thought leaders when it comes to AI in general for me. Every time I go to a trade show, Podium's got something cool to talk about and new. Share a little bit about yourself and then what Podium's got going on.

Jim: Yeah, yeah, thanks, Tersh. Well, first of all, thanks for having me here, this is awesome, really excited to talk to you a little bit. Heard a lot about you over the last 12 months or so since I've really started focusing on home services specifically and applying our platform to that industry, and so I'm just super excited to be here to talk about it. By way of introduction, I'm a Senior Product Director here at Podium. My main responsibility is building our entire AI operating system for home service businesses. We started that journey, started talking about it about 24 months ago, and got building shortly after that. That's kind of my role here. I also handle a few other areas of our product, like our payments product, but that ultimately just serves what we're doing here for HVAC businesses.

Why won't AI take your team's jobs?

Tersh: As far as AI in the trades for contractors, what's your thoughts on that? Because for me, when AI was first introduced to the masses, a lot of people were terrified that it was going to take over everything. The CSRs were worried they were going to get rid of all your CSRs, and technicians hated it. Loved it or hated it. There are products out there, something like Rilla or Siro for the technicians, and I introduced that into one of my businesses, and at first the technicians were super anti. "I don't want Big Brother listening to me" type thing. Then I explained to them how it would help them, and they kind of picked up, they were like, "Okay, I'm cool with that." I like that it helps me learn how to become better without you having to do a ride-along necessarily. So in some instances I'm seeing contractors loving it, and I'm seeing some that are still really hesitant toward it. What's your take on it?

Jim: Yeah, I think that's a great question. I have a lot of thoughts on this. Number one, there's the famous quote, I'm sure you've heard it: you're not going to lose your job to AI, you're going to lose your job to somebody else with AI. I think that holds more true now than ever, and I think there were plenty of AI companies thinking about how to do full human replacement. My personal model, and what I push on our team, is let's focus on turning people at these companies into superhumans with AI, not focus on replacing them. There are pockets where you can talk about human labor efficiency, a good AI CSR working over both voice and text could replace the need for a call center, those kinds of things, but for the most part, nearly all of the AI we're building allows the customers we work with to focus on how they can do more and work on the business rather than doing the tedious stuff that happens day to day. I think that's really important. The user reception to AI and the fear of it up front isn't even necessarily about losing the job, it's "I'm not sure if I want this thing listening to me." I think that's very natural with a new technology, there's a lot of fear about what it might do. What really matters is that the AI you're using actually provides meaningful benefit to you. So when we talk about how you go about picking an AI vendor to work with, for me it always starts with: what is my actual problem I'm trying to solve,

Tersh: Yes.

Jim: and what's my return on investment on buying this AI product? Is it actually going to solve that problem, or is it just a buzzword?

The AHR Expo moment: when "AI" was really just automation

Tersh: There are so many times where I'm guilty of this, it's shiny object syndrome, it's like, "Oh, this is the newest AI product out there," and then sometimes I'll dive into it and I'm like, this isn't AI at all, this is a bunch of automations packaged up as AI. I was giving presentations, man, when AI was first coming mainstream, and I remember at an AHR Expo I was up on stage. I didn't expect it to be as large of a crowd, but it was, I was literally the only one presenting on AI at the event, and it was "how do you use AI in your contracting company." I lost the crowd because I was talking AI, like legit AI, the whole time, and then about three-quarters of the way through my presentation I showed a slide, and it was Zapier. It was literally just Zapier automations, and people perked up and they're like, "This is the AI that I want." It was an aha moment for me, it's like a lot of people don't really want AI, they want automation. So what I found is we can sprinkle some AI into the automation to help, and you made a really good point: remove the monotonous task so you can become better at serving your customers. That's the whole game. When I go to a location and do an onsite visit, nine times out of ten it's your CSRs doing all this clicking and all these other things where they don't need to be doing that. All that stuff can be automated, or we can use some AI tools for that, email follow-ups, SMS follow-ups, those types of things. You can implement one program, you have Podium over here and you're not even using it for your follow-ups, your happy calls and happy texts. I feel like there's so many times where people are afraid AI is going to take their job from them, but in reality it's just like what you said, AI can remove the monotonous task.

Jim: Yeah, what you started with there reminded me of a customer conversation I had very recently. Relatively AI-forward, very AI-curious, and they're using, I don't even remember what software, kind of a legacy software stack. He was telling me he was getting together with other friends who are business owners in similar trades, sitting around a campfire having a couple of beers, playing with Claude or ChatGPT or something, trying to figure out how to use it. You're saying the challenge is we see some cool things and some cool ideas, but we're not sure how to apply it. There's this whole rhetoric in the AI space right now, and we need people like you and me, and honestly the engineers I get to work with every day who are really smart about this stuff and curious about figuring out how to apply things, because it's a technology push, right, this technology is invented and everyone's trying to figure out how to make it valuable. So we need that, but what we're finding in the industry is that the value in AI isn't so much in the large language models themselves or the frontier models, it's in the application layer. Meaning, how is AI actually applied? That's really what automations are: the underlying technology of triggers and outcomes. If you'd talked to your group about triggers and outcomes, they'd probably have been equally confused. You talk about Zapier, that's been really well productized, it's a valuable application designed to fit what triggers I care about, what outcomes I care about, what work does that remove that I was doing manually. That's useful. I think the exact same pattern is happening with AI. So the application layer is where you take the underlying technology of AI and apply it directly to the business workflows, to what needs to happen.

Inside Podium's AI membership coordinator

Jim: A perfect example I'd give is our AI membership coordinator. The story here is kind of interesting. I was early in our journey of building our full operating system, including the FSM. We were spending a lot of time on site with local businesses, seeing how they work. We were on site with an HVAC contractor, sitting with their office manager, seeing how he works every day, watching what he's doing. He had these stacks of papers on his desk, huge stacks. I was like, "What's in those?" And he said, "These are our membership agreements." Each one was a paper membership agreement, with who's the customer, what's their address, what's the state, checklists, have they paid, have they not, that sort of stuff. We were comparing that to other businesses who have maybe their membership agreements already in a field service management software, but then you still see them build a separate Google spreadsheet to work through,

Tersh: who's active and who needs to pay, or whatever.

Jim: Exactly. I was like, "Well, what do you do with these?" And he was like, "Well, when it's time for membership scheduling season, I pick up one paper, I call that phone number, I try and get the customer on the phone. If I get them booked right then, I move them into this pile. If I don't get them booked, I move them into this pile for a callback." You look at that and you go, AI doesn't solve that immediately, you can't go to ChatGPT and say solve this problem for me. But if you wire an AI agent into your field service management software with instructions that say, "When it's membership season, here's my member list who needs to get booked, for each one you need to understand the unique context of their situation, is it the middle of the agreement, is it the end of the agreement and they should probably renew, do you need a card on file, whatever it might be, you should know what filter size we need so when we book it for the technician we mark that on the job," you can get an AI agent to do that. Podium has an AI membership coordinator that puts all that work on autopilot. This customer, with about 1,000 members, was spending weeks, if not months, with one person going through and getting everybody booked. Now he flips on an agent, gives it instructions, and watches his membership appointments get booked all day long. So what can he do? He can focus on how to actually grow. We've kind of been stagnated at about 1,000 members because it's hard to operate that number.

Tersh: It is.

Jim: Instead, let's focus on how we can actually grow our maintenance agreements, how we can train our technicians to be better at retaining customers, better at upselling them on a membership program, better at that customer experience, the part that's still human to human. Just take that tedious work off the plate.

What is an AI agent, really?

Tersh: Absolutely. You mentioned agents, and if you're listening to this podcast, you probably have an idea of an agent. Can you simplify what an agent is for those listening, maybe they're a contractor and they hear the word agent, hear it's a buzzword, but a lot of times people don't understand. I've heard people describe agents as all kinds of different things in layman's terms. Can you break down what that is?

Jim: Yeah, some of the engineers I work with will probably cringe at the way I explain it, but I'll put it the way I think about it. To me, an agent is an AI tool built on top of some sort of large language model, or some sort of set of rules and intelligence, and that can be generic intelligence. Then he's given a goal, given context and a goal, and his actual outcome, his goal is to go achieve that outcome. I'll use an example that's relevant, this membership coordinator agent. He's built on top of general intelligence, this LLM model, and he's given a lot of context. He knows what your calendar looks like, who your technicians are, what their skills are, what their availability is. He knows what the membership agreements actually are, what's in the agreement, who the active members are, what their status is. Then you give him a goal, you say, "I need you to go schedule 12 of my members this week and optimize for route density."

Tersh: That's not—

Jim: He will go off and do that. He'll figure out who's available, when are they available, who do I need to book. He'll plan how he's going to reach out to those members. You also have to give agents tools, the things he can actually do in the platform. For our membership coordinator agent, that means he needs to be able to send a text message to a customer, track if they followed up, respond to them, answer their questions. He needs to be able to book a job on the calendar, assign the right technician and customer once we've confirmed a time. He needs to be able to give answers back, report what's happened, to power a reporting dashboard to keep track of where things are.

Tersh: Do you have the ability to do outbound calls using AI voice?

Where does voice AI actually belong?

Jim: We are working on outbound calling right now. Candidly, in the HVAC space, I'm seeing this interesting thing going on in the industry, if I can just zoom out a little.

Tersh: Yeah, absolutely. Right.

Jim: I think voice AI has a real place, I want to start there, I'm not saying it's a bad thing. What we have seen over the last 12 to 18 months that we've had our voice AI agents out there is that there's a time and place for voice AI, and a time and place where it doesn't necessarily make sense. If you're a small business and you don't have reliable after-hours call handling, or your team is already on the phone and it's a call overflow scenario, there's very little downside in turning on a voice AI agent to pick those up and make sure you don't lose that lead. Feeding the leads is most important. That said, we've seen pockets across the United States, based on consumers, nothing to do with the technology, where consumers are saying, "I don't want to talk to an AI agent, I want to talk to a human." So what we're finding and pivoting toward is let's set up the voice AI to apply when it should, and not have it always catch everything. If a business says, "I just want you for call overflow and after-hours," that's great, and the value is there. But there's still this uncanny valley to get over, where voice agents especially can feel weird and could do damage to your brand if that's not how you want to operate. We're totally supportive of that, because the value is still there, you can still handle things after. That's kind of where I'm at with outbound calling, there's a time and place for it. There's a lot of legal stuff to work through to make sure things are handled the correct way, and ultimately what matters to me is the ROI has to be there, and part of the ROI equation is protecting the brand reputation.

Hang-up rates, sentiment and the PhoneTap data

Tersh: Absolutely, yeah, 100%. With PhoneTap, I have two co-founders, and we analyze phone calls and basically save the leads businesses are already paying for. One thing I've noticed with AI voice companies is, if they announce they're an AI voice whenever the client picks up or calls in, it's a drastic drop-off, their sentiment analysis is really low to start, so it has to battle the negative sentiment, a lot of times they're asking for a human agent or human CSR. What a lot of people aren't paying attention to is the hang-up rate. As soon as they realize it's an AI voice, your conversion rate is whatever it is when you're actually talking to them, but what's your hang-up rate when they realize it's an AI voice? The companies that aren't announcing they're an AI voice are, I'm thinking, like 20% better at converting using that AI voice. There are other factors in there, but you're exactly right, there are definitely geographic areas, like middle to southern Florida where a lot of people go to retire, they do not like AI voice agents, at least so far what I've seen. I'm sure there are some voice agents that might work, but based on what I've seen, and we've analyzed hundreds of thousands of phone calls, I was actually talking to a client of mine yesterday and he said his AI voice is doing worse than his worst CSR. I wanted to dive into the details of that, because I'm curious when are they actually picking up, is it after hours, because you should take that into consideration when you're judging it against the CSR. Or I believe personally that the mortgage companies, or the lending companies, are ruining it for outbound AI voice, because I must get 15 to 20 AI outbound calls a day. It's so common, and I've talked to so many other people and they're getting these AI voice calls, and if you do happen to pick up the phone, they transfer you to someone else as soon as you pick up. It's ruining it for everybody.

How does Podium's AI Studio train and improve voice agents?

Jim: That's interesting. This makes me think about a study we ran recently. One of my team members led the study, and we were encountering similar things, hang-up rate might be bad. We tested a bunch of different things. One of the things we tested, in addition to AI announcing itself, we tested having the business owner record a pre-recorded message that says, "Hey, I'm Bob, I'm not available right now, but my AI agent, I'm going to hand you to Jessica and she can handle the issue and get you booked, we would love your business," something like that, whatever the message is. What's interesting, to your point, is it seems to be so regional, or so dependent on the brand and who your target customer base is, what works. There's no one-size-fits-all. But what we found in these tests is we've gotten to a point now with our voice AI that hang-up rate is actually no longer the problem. We were really nervous, we were looking at hang-up rate, and I think it's like a 20% hang-up rate on inbound calls. What we did in this test is we sampled calls where our AI employee picked up versus calls where a human picked up, and we found the hang-up rate was the same. We got it to a point where AI employees were not losing any more than humans were. That made us realize a lot of these hang-ups are actually, "Oops, I called the wrong number," or things of that nature, or just straight-up spam calls where the business hangs up too. That was actually kind of an interesting finding we had. Can I make a little plug for Podium on this? Because everyone has such a different situation, between brand, region, customer base, all these things, this is what's great about our AI Studio. It's where you go to instruct, train, test, and give feedback to your AI agents. There's a dashboard that shows all your agents, metrics on how they're performing, that's one piece. What's really exciting is if you see a call that happened, you can immediately, inline, give feedback to the AI agent, "Hey, you said this thing wrong," or "I want you to change how you introduce, introduce it this way." Behind the scenes, because this is really a system of agents, there's a trainer agent that picks up that feedback, goes and looks at the playbooks and policies the voice agent was running on, modifies those, and then says, "Why don't you call me and pretend to be a customer, and we'll run it again," or "Go show me what you would have said in that actual call, had you been working on this new playbook."

Tersh: Oh, that's cool.

Jim: It's wild how well it works. It allows people to customize and train, and that's so important. When voice AI first came out, and many vendors still work this way, when someone gives feedback they have to get an engineer to go try and update it. The amount of customization needed is so vast, you have to build products like this that allow people to customize to get it to work the way they want.

Tersh: That's the thing about it. Whenever we talk to clients, we wanted to make sure we were very customizable for that reason, because no two businesses are run the exact same way. If you have an AI voice that only answers one particular way, or you have to painstakingly manually update every single response to any possible scenario, that's such a turn-off, that's so much work for a business owner trying to implement something so they can do less work. That's really cool though, I love that you're able to do that, and it works the way so many people thought AI was going to work when it first came out. We heard AI's here and immediately we're thinking I, Robot, good or bad, when you think of an agent sometimes the thought was this agent was actually a robot working behind a computer typing things out, and other people are like, "Oh, this is going to be I, Robot, they're going to take over the world." I don't think it's going to be there yet.

Personal AI agents and owning your own data

Tersh: Let me transition into your opinion on this. I talk to a lot of people, and it's becoming more popular of a conversation, people starting to have their own personal servers, housing their intellectual property locally and tapping into that. Have you heard people talking about that, like a business or a tonality type thing?

Jim: A little bit.

Tersh: It is very new, it's a newer thought concept as far as the masses go.

Jim: Yeah, what are you thinking about?

Tersh: I have it personally. I have Carl, my local agent and server, and every conversation I have immediately loads into Carl, and every conversation a technician has immediately loads into Carl, and it can be used for coaching, because the thing is we can gather all this information, but what do we do with it after we've gathered it? Having everything go into Carl, you can get a weekly summary, "Hey, these are conversations you had." I actually have a wiki page, so it's like, "Do you want me to update your wiki page," to say this is your thought process on this concept now, because it recognizes the trends in my conversations and says, "You used to think it was this, you used to have this thought process, now you're trending in this direction, do you want me to update your wiki page?" Then whenever it writes my emails for me, it references my wiki page for the tonality. I have this set up for each of my businesses as well.

Jim: Yeah, I probably don't have a lot to say specifically about where you're hosting that data, your private server, but I think there's massive benefit to people where the value is there to set up their own agents. I've got my own personal assistant agent, I've given him a name, I haven't personified him, but my personal assistant, I'm primarily using Codex right now, jumping between Claude Code and Codex, and my personal assistant repo in Codex, he knows everything. He's wired into Slack, our internal communication tool, connected into my Google Workspace, he's got access to call recordings I have, every conversation I have. What I think is so impressive is the amount of work he can actually do for me when I sometimes just word-vomit my random thoughts to him on my drive home, and he can go do things. I will say, every once in a while, when you're doing this a lot, my CEO can still detect it, and he'll ask me, "Did your personal assistant send me that, or did you actually type it?" There are still kind of those things, but I think these are super important, and context, context is really the game. Can you have context across everything going on, can you train and give it instructions on what to do? I think those things have massive value.

Tersh: Yeah, I agree. I don't know the ins and outs of Podium, obviously don't know the back end, but I feel like that's sort of what these agents are doing, like your AI voice, you can explain to them, "No, this is the way I would do it in that scenario." There are times where I've listened to AI voices and I'm like, that's not right. And what's right? I don't know, but that's not it. There are times where I can't tell you how to fix it, but I know that's not right.

Jim: Yep, yeah, I think we see that all the time. It's hard to distill out what it should be doing, but it's easy to say that's not right and this would be better. That's where, going back to the analysis we did of voice AI calls versus human calls, what worked better and where things dropped off in the funnel, honestly, we find some of the best ways to train these models to be really effective is to focus on how the best people are doing it today.

Tersh: Yeah, absolutely. It's great to have that database. I never thought that was going to be the case with PhoneTap, I thought PhoneTap was just going to save leads that were missing, a bookable lead that wasn't booked, send a notification through Slack, let's book this lead, you already paid for it. But what I'm finding is we're learning how the people doing a really good job at booking, really high rates, how are they talking, what's making that work for them that isn't working for others. That's just a byproduct I didn't realize was going to happen, which is really cool. We're at the end of our time together, but I'd love to hear about what's in the pipeline for Podium, what sort of AI tools are out there, because a lot of people don't know how advanced things are with Podium.

Speed to lead and the birth of Podium's AI employees

Jim: Yeah, thanks, I'd love to talk about that. First, just to set context on the story about why we are where we are: well over two years ago we started launching our first AI agents, with the primary use case of handling leads coming into your business, starting over text and expanding to voice. We were doing that at the time when we thought of ourselves as a communication and lead management platform across a variety of verticals, auto, med spa, HVAC, the verticals Podium likes to focus on, all with similar needs around a large influx of leads coming in every day, and speed to lead is the number one thing. If you don't respond in five minutes, the chance of winning the lead drops by 90%, maybe even higher in HVAC, we all know we're just going to call the next company until someone picks up. So speed to lead, super important, that's why we started building AI employees, always on, always behave the same way, never have a bad day. We started there. How I like to explain this: a decent AI agent, at least at the time, is one that can answer the phone and say, "Hey Tersh, totally understand your problem with your furnace not turning on, that sounds pretty bad, I'll send this off to a human and someone will reach out to get you booked on the calendar." A better AI agent is one that says, "Hey Tersh, understand the problem, totally get it's cold outside, this is kind of an emergency, Jim is on call right now, he could be at your house in 30 minutes, do you want me to send him?" The delta in those two things is really what led us to building our full operating system. The delta is now your AI agent can't just simply understand and write a note and send it to a human, your AI agent has to understand who's on call, does he have the right skills for the work, when is he available tonight, what's our dispatch fee so I can communicate that with the customer and get them aligned, what's the actual drive time to get there, all these things become important, and is it a type of job we actually do. In order to get there, that AI agent has to have access, ideally native access, to all of that information about how your business runs, so it can have the right context and do the right tasks.

Meet Operator, Podium's back-office AI agent

Jim: So we built a full operating system platform that's not just the front office communication layer anymore. It's a full field service management software under the hood that can manage technicians, price book, membership programs, a technician mobile app to go actually do the work, track the jobs, inventory, estimates, payments, the entire stack. Where we're at, there's three main AI agents to talk about. Number one, we've talked a lot about the CSR already, and I already explained the membership coordinator, those are the first two. The third one, honestly the most exciting to me, is what we call the Podium Operator. The way to think about Operator is, imagine you had another employee or assistant in your office who could do work for you when you need them to. The way we built Operator, he basically has two things. He has skills, as in, "When someone's asking about this or wanting to do this, I know what to do," a very human-readable playbook of how I should operate in these scenarios. Then he has tools and access, meaning he has access to all of the different data types in Podium and other information you might connect him to. What's Operator, in a way it's almost hard to explain because the use cases are so widely varied. We see business owners using Operator for things like, "Run a daily scheduled report that tells me how the business did yesterday across these 10 different categories, and only surface to me the areas where we're not performing well, and don't only tell me what didn't go well, actually propose insights and actions I can take to fix it."

Tersh: And we don't want to just see the problems, we want the solutions to the problems as well.

Jim: Exactly, and in a way it goes back to your automation thing. We know reporting, everyone's got reporting they use. Most reporting, up until the world of AI, has been, "Hey, you can go create a report and check it anytime you want." Better reporting is one that can actually produce insights: is this property drop actually meaningful or interesting, do I need to be paying attention to it or is this just the season? And one that can actually take action on that, recommend what to do with those insights.

Turning reporting into marketing ROI and live dashboards

Tersh: We should have started the whole podcast with this, because I have a gazillion questions, but we don't have time for all of them. I do want to ask, is it connected to marketing, so it could essentially say, "Hey, you should ramp up your Google PPC or your LSA," because that's where you're getting the best leads, because it's analyzing the marketing spend as well?

Jim: Yep, exactly, exactly. You could plug in what your marketing spend is. If that spend is actually integrated directly into Podium, if you're using some social campaigns or Google LSA, you can automatically have it calculate what the ROI is on that marketing campaign, how it's doing, and it could produce that for you.

Tersh: That's really cool. What about something like, say you want a three-day call board or something like that, does it have the ability to do something like that?

Jim: Yep, are you talking in terms of actually generating the views and everything, or—

Tersh: Sure.

Jim: We've built a lot in Operator where you can actually generate live, rich user interfaces on the fly. That's another thing that's wild with AI, it's getting more into the coding space, but AI is actually really good at writing code, and on a prompt, if you ask for a report, it can generate a visually good-looking report that shows what's going on.

Tersh: Not just an Excel spreadsheet that goes to your email which you never look at again.

Jim: Yep, exactly, exactly. So he can definitely do that.

Tersh: That is so cool, man. That's really exciting. The world of AI doesn't have to be scary, it doesn't have to be overwhelming, it's when you use these tools properly and use them to their fullest, too. If you're going to get Podium, don't just sign up for it and have it send out an email reminder every couple of months. Use it to its fullest potential, it's really a great investment.

How did Operator fix a technician job-notes problem?

Jim: Yeah, there's one interesting story I love, because it's so simple and kind of transient, but I think it speaks to this. So often we think about software as, "Hey, you build these whole systems, you get them all configured," and you're only going to worry about building and setting up the things that are evergreen and always necessary. Talking to one of our customers who's been using this Podium Operator for a while, loving it more and more, finding different use cases, he had a very pointed issue in his business where he noticed his technicians weren't leaving notes on their jobs. That's not something where, if we go build a technician job-note-completion dashboard, that's going to take a lot of time to set up, and then it's probably going to be something we look at for a couple of weeks, and then—

Tersh: it wears off.

Jim: He told Operator, "I have this problem, my technicians are not leaving notes on their jobs." That's literally all he said, and Operator said, "Okay, cool, here's how I'm going to help you fix that. I'm going to set up a daily report that you're going to get every day, and your technician leaders are going to get that report every day. I'm going to look over the last couple of days and show you who's actually been leaving their job notes and who's not." The extension of that is you can say, "When someone's not actually leaving job notes, Operator, go text that technician and ask them what happened on the job they forgot to leave notes on." He ran this internal campaign with his team, and I'm sure he gave some sort of rewards to people performing best, and he was able to stand up not just a tracker to go hound people, but an entire process to keep track of and improve the team's performance.

Tersh: There are so many times where I've done that, where I've built out a report that was meant, I had every intention of holding people accountable using the report's information, and then I'd do it once or twice, but then I have to remember to look at the email, or remember to look at the report, and then go through and painstakingly send a text message out. But the fact that it does all of this automated is the next level. That's where you're using AI smart, and not as a shiny object, "Oh, this is the new technology that's AI-enabled or whatever."

Jim: Yeah, yeah, exactly.

Tersh: Dude, this has been a great conversation, we could talk for hours, I know. For those who are interested in Podium, what's the best way for them to learn more?

Jim: homeservices.podium.com, great place to get started, to learn about our AI operating system, and from there we'd love to talk directly, our lines are open, and we're just really focused on helping home service businesses make more money, be more effective, and win every day.

Tersh: I think we have a lot here. Cool, Jim, appreciate you.

Jim: Thanks a lot, Tersh.

Tersh: Thank you. See you.