AI Employee: How Home Service Businesses Are Scaling Without More Staff

What's happening at AHR Expo 2026?

Dennis Steele: You can't get them upset, ever. So our AI employee has basically two modalities: able to text, able to talk on the phone, able to book appointments straight into the calendar. The other intangible is Larry's not going to get upset. He handled it super clean and ended up with a booking. It was pretty amazing.

Tersh Blissett: Hello everyone out there. I hope you're having a wonderful day. If you're listening to this podcast, you may hear some background noise, and that's because we're live here at AHR 2026. What's up, dude?

Tersh: I have one of the co-founders of... why are you looking so surprised? I love this audience. We have one of the co-founders of Podium, and I'm super excited to talk about the origin story, AI in general, agents, how AI is changing things. This is one of my favorite companies — not to mention I have a lot of great friends who work there. I love how forward-thinking you are.

Tersh: I was talking to Bryce — oh man, it's probably been a year ago maybe — and we were talking about AI agents. He's like, "Dude, you've got to check this out." In Bryce fashion: "You've got to check this out, dude, you've got to check it out." And I'm like, what is it? "You just don't understand how awesome this is." I was like, all right, what you got? And he showed me, and I was like, okay, that's pretty awesome. So I'm excited. He was like, "It's so good, you should try and break it." And I was like, yeah, I should try and break it. I did not try and break it. But I do love the functionality. I love how forward-thinking you are in general as far as AI goes, and I had no idea you had so many team members. With that being said, do you want to share who you are and how you even got started doing this?

How did Dennis Steele get started with Podium?

Dennis: Yeah, so my name is Dennis Steele. I'm the co-founder of Podium, and we started back in 2014. So it's been a minute, and it's such an exciting time right now with AI and AI agents. I never would have expected, when we started the business twelve years ago, that we'd be in this position — almost reinventing ourselves as a business and a product. But that's where we are. The technology has shifted that much.

Tersh: How crazy is it, the things AI has changed? Because when I first thought of AI, I was thinking automation — that's where my head was at.

Dennis: Yeah, it's automation. And then it turned into real AI.

Tersh: And I'm like, okay, this is cool. Your company is large — I've always heard it's great to stay small because you can pivot. How do you transition a big team like you've got?

Dennis: It's tough. It's tough. But the biggest thing is — everybody goes through this. When they first start playing with AI, they see the power of it and think, oh, this can do some automation for me, it can answer some questions for me. The big breakthrough came when we saw the power of agents. For those in the audience who don't know what an agent is —

What is an AI agent, really?

Dennis: An AI agent is when you're leveraging the AI model and giving it a set of tasks, along with enough context, to go complete those tasks. That's what, the past two or three years, when I presented on stage and said we're going to talk about how AI and automation are going to help the trades, people would come in and say, "I want some of that AI." But what they really wanted was automation — what they actually wanted was agents. It just wasn't until recently that agents even became a possibility.

What's the shift from AI assistance to AI employees?

Dennis: Initially with AI, what you'd look at is, oh, I can auto-complete something, or I might be able to better determine how to set off a trigger in software — it's a really fancy chatbot. Where we've come now is we're seeing agents do human-level work across multiple tasks, consistently, over time. That's a threshold we've crossed over the last year or so, and that's where it transforms: instead of just assisting you or being a co-pilot, you can set up a system where that agent can actually go make you money. Because the reason you hire a person, a CSR —

Tersh: You have to have ROI.

Dennis: Yeah, they have to have ROI. You're paying an entire salary — how does that work out economically? Well, they're going to produce additional revenue for your business, or help you support additional revenue. And now what we're seeing with these AI agents is that they're giving businesses the capability to do that.

Tersh: And are you building out agents for clients, or are you using agents internally?

Dennis: It's kind of interesting, and I can get into a little of the backstory — for the first ten years of the business, we were building a lot of the foundation that our agents use today. We were a communication platform. We started out helping HVAC companies with reviews; it was a basic point solution. My co-founder and I saw a shift in technology, similar to the shift we're seeing now with AI, and that was Google putting reviews online for businesses. That shifted everything — it changed where all the leads were going and how businesses were getting leads. The businesses that adapted to that shift started to make more money and stay ahead of the competition, and the ones that didn't struggled. We're seeing a similar shift now with AI. It's changing the game, and it's a little scary — the ones who are leveraging it are able to get additional revenue and figure out how to automate their business and find more efficiency, and the ones that aren't are struggling to keep up.

How is AI changing customer expectations?

Tersh: Do you think that if people are purposely ignoring AI, they'll suffer? Will they go out of business, or will they just have to do business completely differently?

Dennis: I think it'll show up in two different ways. One, consumer expectations are shifting. The more businesses use it — it's like the first time you use Uber, it's such a smooth experience, you never really want to go away from that. Once you work with a business that can support you 24/7 and have the answers you need whenever you want, it's hard to go back to playing phone tag with a business over the course of a couple of weeks trying to get the answers you need. So consumer expectations are going to increase, because as businesses adopt AI, you can have better service as a business.

Tersh: Are people going to become more lenient and understanding toward agents, or are we going to have to make the agents a lot more human?

Dennis: Both. If you're working with an agent that's not great, it'll hang up really fast or whatever — so we're going to have to make that a lot better. At the same time, it's a balance, because I don't think most people are going to want to go through their entire experience just dealing with an agent. There's a lot of human aspect that makes an experience really good.

Tersh: Have you ever experienced someone coming to you and saying, "I was talking to this person," and you're like, wait, that's not a person — that's a voice agent? And then, well, actually I'm sure it was a person.

Dennis: Yeah, that was definitely a person — people are so confident about that, that I've had to show them the recording and the transcript from our AI voice agent to prove that it wasn't a human they spoke with. We're hearing that all the time.

Are customers building real relationships with their AI agent?

Dennis: I was just talking with one of our partners, and their customers are like, "Hey, we call our AI employee Larry" — as a business, you can call Podium's AI employee whatever you want. But customers develop a relationship: "I worked with Larry, he was great, he helped me out." We've seen reviews with Larry's name in them. Larry did a phenomenal job taking my call. We hear that all the time. Part of it is he's thorough, he's consistent, he's courteous — you can't get him upset, ever. That's the best part.

Tersh: I've told people before, if you're not sure whether that was an AI agent or not, start arguing with it. If it doesn't argue back no matter what you say, it's just chill as a cucumber.

What happened with the angry driver who got booked anyway?

Dennis: There's a pretty good story about this. We have a customer, Aire Design, in Utah — about a ten-truck HVAC business. A random guy gets cut off by one of their techs at an intersection; the tech blew the red light and the guy gave him the one-finger wave. So the guy's mad — he texts in to the business: "Hey, one of your techs just cut me off at the intersection and blew the red light." So upset. Larry takes over: "Hey, so sorry, we care about tech safety" — handled the whole string perfectly. And the guy is so impressed he ends up saying, "Hey, why don't you guys come out to the house and give me a quote?" That's a booked appointment, and Larry was able to book it straight into their CRM, handled it all.

Dennis: So our AI employee basically has two modalities: able to text, able to talk on the phone, able to book appointments straight into the calendar. But the other intangible is Larry's not going to get upset — he handled it super clean and ended up with a booking. It was pretty amazing. And the other thing is Larry can be simultaneously handling an irate customer while another customer calls in and you're booking a service call there.

How does AI handle unlimited call volume?

Dennis: When you look at the way agents are really transforming the business, it comes with scale. You can turn Larry on and say, hey, after two phone calls during our busy hours, Larry can take phone call three, four, five, six — up to an unlimited number of phone calls. You're never going to hit capacity. That's hard for an HVAC company — you're trying to staff to capacity, figuring out how many CSRs you need at any given time. The biggest concern used to be, what happens when it slows down? I don't want to lay off staff. But if you have something like Larry, when things slow down, he just answers fewer calls.

Dennis: After hours, over the weekends, middle of the night — if you could hire a CSR and say, "We're okay during business hours, but I actually want you to answer the phone at the hardest times: after hours, weekends, holidays, 24/7," and not sound like you're falling asleep or just woke up — that's where Larry comes in, and he performs extremely well.

Is outbound AI calling on the way?

Tersh: So how about outbounding — is that a possibility? I know there's a lot of stigma and regulation around it. Some people love the idea because they hate outbounding in general, but I know at one point there were regulations against using AI voice to outbound.

Dennis: I will tell you my voicemail box would beg to differ — every single day there are like twelve thousand messages in my inbox of AI voice outbounding.

Tersh: Are you able to do that with Larry, or is that something you'd rather hold off on until the legislation shakes out?

Dennis: No, outbounding is big, and we're going to be releasing it soon — it's going to be a money maker. A lot of the time it's not going to be just straight robocalling. You have customers, you have memberships, you have a membership program, and this is where we can use the AI agent to help fulfill and reach out and generate business and fulfill your calendar. When we talk to our customers, a big piece is managing and fulfilling your membership program — sometimes it's a full-time job. So we can upload all that information. We're splitting it into two products: a membership management agent that can help you go fulfill, and you can pair outbounding with that to help fill the calendar — the outbound agent can analyze your ServiceTitan account and say, hey, this account has an open estimate that needs follow-up, here are all the open estimates, reach out to them and see if you can close those deals.

How can AI use customer data to generate new jobs?

Dennis: We're pulling all the data in, we're looking at the opportunities for you. The initial phase of Larry has been all about answering phone calls and making sure no leads go to your competitor. The next phase is: how do we monetize your data? Data's a big topic — everybody's talking about data, but how do you monetize it? Having an agent, and the power of AI on the analytics side, to determine when's the best time to reach out, when am I going to fulfill, instead of just blasting out messages and getting shamed on. You can have the agent reach out at the right time, look at the calendar, and know, I'm going to fill the calendar over the next two weeks with the customers who are in the best spot to help us out.

Tersh: Are you able to analyze the lifetime value of the client and figure out, okay, these people have reached the edge of their lifetime value, so they're not as high a priority to reach out to, versus these over here who are right in the prime of their full lifetime value?

Dennis: A lot of that's going to be done with AI analytics — you don't have to have a data scientist figure that stuff out. We want to get to the point where the best thing about these agents is you can just speak to them in plain English. You say, hey, analyze the lifetime value of my customers, I want to figure out the best campaign to run — that's where we want to get to, that's something we'll be coming out with. Right now we're working on the lowest-hanging fruit: cover the phones, answer your calls, 24/7. Nobody wants to do that as a business owner. I hired outbounders who were commission-only, and they didn't do an amazing job, but my wife did it, and she had a completely different passion because she owns the business, so people could hear that inner voice — she really cares about those people. The outbounders we had didn't have that same inflection in their voice, so everybody hates doing it because you get told no constantly. There are no feelings involved with an AI agent.

How do you customize what your AI agent says?

Dennis: What we've done, and this really separates Podium, is we've built a really complex customization piece into our agents. If Larry says something you don't like, you just tell him — you don't have to go in and rebuild the prompt. We have a thing called AI Studio. You go in and say, "Hey Larry, don't ever say it that way again, say it this way," and he writes that down into what we call a playbook. When we onboard you, you build out this playbook: incoming calls, after-hours, all these different scenarios. This customization is really important because every business is different.

Tersh: Here's the crazy thing for me — my CSR Julie will be on the phone with a customer way longer than she should. I'm like, let's wrap this up, let's finish it up, and she'll be asking, "So how's your mom, how is she?" and I'm like, I don't care about their mom — but she had surgery on her leg — again, I don't care, let's book a service call. She's like, yeah, but I want to check on their family.

Dennis: Those are the little nuances — with AI Studio you can just say that. "Hey Larry, you're being too wordy, you're asking too many personal questions, pull it back a little bit." He'll write that customization, then play the conversation back to you and say, "How about I handle it like this next time?"

Tersh: How long does that take?

Dennis: That's instant — you're working with the model right there.

What happens when the AI goes off script?

Dennis: If you wanted to pick something safe — a service technician calls and says, "Hey, go listen to that call." I go listen to it, it's three and a half minutes long, and I say, all right, I've listened to it. The technician says, no you didn't. And I'm like, well, you're right, I listened to two minutes of it. So I go back and listen to the remaining minute and a half, and right about the three-minute mark, Larry — or maybe it was someone else — starts singing opera back to the homeowner. And nobody says anything. The homeowner doesn't acknowledge it at all — she's just like, "Okay, I'll see you then." My technician dances around it without ever actually asking her why she didn't say anything to our voice agent when it started singing to her. So I called her, and when I spoke to her on the phone, she said, "I figured y'all were just trying out singing to people." You'll find these scenarios where the AI kind of goes off — she was full-on opera. We haven't seen the opera singing since, but that was a good one, I've got to find that recording somewhere so I can start using it in my presentations.

How did Podium's review platform get started?

Tersh: Tell us about the initial startup of Podium — I don't think I've ever heard the origin story.

Dennis: Eric, my co-founder, is our CEO, and we started the business together. We knew each other in college. He grew up working at his dad's tire shop, and he saw a local blue-collar business doing really well, but the technology it had to succeed as a business just wasn't great — most of these systems at that time weren't. I was working for a local marketing company, and we were helping businesses deal with this huge shift — Google had just come out with reviews overnight. It basically shifted from the business paying the most advertising dollars on the Yellow Pages or a billboard and getting the leads, to overnight, all the leads going to the highest-rated business. It didn't matter if the first three letters of your name were all A's — Triple-A Service — it didn't matter anymore, the leads were coming through a different channel now. That shift spurred this moment of innovation where we said, hey, we can help these businesses get reviews. That was the target to start with — we didn't have any experience, we didn't have any money, so we said, what's one problem we can solve, let's build a little app and try to sell it. That was our reviews product. Here we are, 1,100 employees later. Part of it was luck, but we hit the right timing of a shift in technology that was really valuable, and that fueled the growth.

Dennis: So we raised some money, started growing the company, and that's when we started seeing the power of communication via text message. We built out the full comms platform over the next ten years. What we found with AI and agents is there's a lot of AI products out there, but I think what separates us — and this is by luck — is that for the first ten years, everything we built set this up. If you want your AI agent to make a phone call, we spent three years and fifty million dollars building out a phone system. You want your AI agent to text your customers — we spent years building out our text messaging inbox and designing it for HVAC companies. So we plugged them in, and we already had the automations: the marketing automations, the payment piece. We had a lot of these features we were able to package and give to Larry, and it made him really powerful right out of the gate.

Tersh: Did you A/B test a male voice versus a female voice?

Dennis: That's a great question, and we're trying to find the data. I think we'll find patterns that might indicate what has higher conversion — we haven't found anything definitive yet, but we do have some customers say, "I switched to a female voice and we're seeing higher conversion rates." It might also vary by region — different accents and things like that. We're working on some voice customization features we'll be coming out with.

How should you onboard and train your AI employee?

Tersh: So what's the future? What's next?

Dennis: It's hard to keep up these days. Even at Podium, with 1,100 employees, we're pushing our teams to be on the edge of this stuff every single day. Even this weekend I was at home texting my co-founder, and we're coding on our laptops — we're not engineers, but with AI you can do that now. We're playing with the latest AI models and tools. Some things came out even in the last week that we think completely change not only the way we work as an 1,100-person tech company, but what our job is: to package that in a way our customers can leverage, as soon as possible. We're taking these advancements in AI on a daily, weekly basis back to the product team and asking, how does this change the game? How are we going to write code faster, ship product faster, and then how do we take this to our customers and help them make more money with it — and make it so they understand it enough to implement it, instead of it sitting on the proverbial digital bookshelf collecting dust. That's the art of what we try to have our teams do: take this super advanced technology that we're figuring out weekly and package it so it's plug-and-play for the industry.

Dennis: The thing we do notice is that with the agents, the more context you give them, the better they do. The more playbooks you can give them — this is how we do this, this is how we run this process, this is how we do after-hours — the more you give them, the better they're going to perform. We've basically framed it as: you're going to have to onboard your AI employee just like you onboard a human employee. The mandate we give our team is, one, the agent has to be the best it possibly can out of the box — if someone just pushes play, we want it to be that good. But then you have to have the customization for every time it messes up, so I can tweak it toward being fully customized and always knowing how we do things.

Tersh: Let me ask about this, because it teeters on intellectual-property territory. If you're in a best-practice group and they say, this is how you should do your CSR playbook, and I'm ignorant and just copy and paste it — is that a hundred percent okay, since that's someone else's intellectual property? Or if I've paid them to coach me, do I now own that IP?

Dennis: It's interesting, because the way the models work, they're taking that a lot of times as a reference — the agent has dozens and dozens of other playbooks to reference, so the output is, I took those best practices, I also analyzed the other playbooks you've given me to tell me how to do my job at an HVAC company, and the output is a combination of everything. So there's less of a concern there.

How do you build strong company culture at scale?

Tersh: I've been to Podium headquarters and experienced the employees and team members — everybody's great to be around, just positive people. I especially loved the beanbag chairs.

Dennis: I wish there were beanbag chairs.

Tersh: But how do you develop that kind of culture when you have so many team members? You have to hire on purpose — for us, if I've got five technicians, I'm not taking just anybody.

Dennis: You have to be intentional about hiring — it's one of the hardest things, and it has the biggest impact. What we've learned over the years is the more direct you can be about your culture with your prospects, the better. There was a phase when it was really hard to hire in tech, where we tried to be everything for everyone, but then it's hard to actually have a culture that means anything. So we learned to be direct about our culture: this is who we are, we're going after something ambitious, we have a big mission to help these businesses be successful, and make sure you're finding people who are bought into that. It might take a little longer to find people, but it makes a big difference.

Tersh: Was this all trial and error on your behalf, or did you go somewhere — like people who go to Zappos and say, they have a really cool culture, let's mimic it?

Dennis: A combination, but a lot of trial and error. We had a phase where we looked around the team and thought, man, it's hard right now, we might not have everybody completely bought into our mission. Then he brings somebody like Bryce on, who's just a cheerleader for everybody. You have those locker-room leaders who influence the culture in such a positive way and really bring the best out of the team — those people are incredibly important. People want to be mission-driven, they want to do important things and be aligned to that mission. So that's been the best thing for us: be direct about it, and find people who align, so everyone on the team is aligned.

Tersh: Is there anything I've forgotten to ask you? How many kids do you have?

Dennis: Four kids. My oldest is ten, she's a dancer; my son is twelve months and three days younger, he's into everything sports, my only son; and then two other little girls, seven and four, who are also dancers. We're in the thick of dance competitions right now.

Tersh: I'm the exact opposite — three boys and one girl, fourteen months apart for the oldest two, thirteen months apart for the youngest two.

Tersh: I appreciate you hanging out with us. If anybody wants to learn more about Podium, and they've been living under a rock and never heard of it, what's the best way?

Dennis: Podium.com. It's easy — go check it out, or get with somebody like Bryce, he'll tell you how it is, he's straightforward. It's hard because there's a lot of AI hype out there, and it's somewhat deserving of it, honestly — once you see what it can do and what it unlocks, and what we're seeing happen for our customers, turning it on and making more money within a week, it's hard not to hype it.

What should service businesses know about AI safety and voice cloning?

Tersh: With that being said — and I absolutely will not end on a negative, this isn't necessarily negative, more so security — what are your thoughts on making sure you stay secure using AI?

Dennis: I think it's going to be more and more important as we go into this. How do we protect ourselves, protect our families — because you have tools like ElevenLabs, AI voice synthesis, where you can mimic anybody's voice with a small snippet of video. It's getting wild. You can clone a voice now within seconds and it's indistinguishable, down to the small breaths. We're all going to have to be careful and maneuver this together. Part of our mission as a company is to help figure out not only how to use it to make money, but how to be safe with it — because it's not slowing down, and it's going to require us to stay on top of it as it comes through the pipeline.

Tersh: Does anybody have any questions out in the audience? I know I did an amazing job as a host and asked a lot of great questions.

Tersh: Cool, man, I appreciate you hanging out with us.

Dennis: Yeah. Thanks. Thank you guys. Thanks everybody.

Tersh: Thank you for listening to this episode of Service Business Mastery. Now that you're equipped with essential business advice from this conversation, you're one step closer to becoming the successful owner of your dreams. If this episode has been helpful to your business journey, don't forget to subscribe to the show, leave a rating, and share it with other owners as well. Visit servicebusinessmastery.com to learn more.