Brad Scruggs of ZyraTalk Reveals AI Secrets for Never Missing Calls | Home Service Business Tips

Welcome back to the show. I'm your host, Phil Risher. And on today's episode, we're going to be discussing how to use conversational AI to book more jobs for your home service business with Brad Scruggs. Brad, thanks so much for joining us today, man.

Yeah, I'm excited to be on. I've been watching some of the podcasts and some of the other hosts that you've had on. So, yeah, excited to be a part of this.

Yeah, for sure. So you're the founder and CMO at Zero Talk. Can you tell us, like, what is Zero Talk? How did they even come. Come alive? What is it?

Yeah, I can give a bit of a background and kind of where we started, we've been in the home service space for about eight years now. We actually started with a different brand called Great Pros. And we were essentially, we were doing marketing, we were generating leads and we had a network of contractors from H vac Plumbing Electrical just to handyman who were buying leads from us. And we were spending all this money on Google Ads and generating thousands of leads every single month. We had a contractor network of about a thousand contractors and across different verticals. And we always wanted to challenge online forms. And all the other players in the space were using forms to capture leads. HomeAdvisor, Angie's List, Porch, all these other lead generators. So we were like, man, if we could do something different just to increase our online conversions even by a fraction of a percent, that would be huge for us. So we built an automated chat platform for ourselves. It wasn't complete AI at the time. We, we were using IBM, Watson to analyze intent. So we were analyzing if people were asking about pricing, scheduling, some different scenarios, and we had different conversation flows built for all of this. But as of the last two years, with the, with the advancements of OpenAI, ChatGPT and all these other LLMs, we've made it full AI. And you know, there's a bunch of LLMs out there. We tested about a dozen ChatGPT. They're miles ahead. So that's the brain behind some of the AI that we're using. And then the name of our company, Zyra Talk. We've always wanted to get into the phone. So when we started Zyra about seven years ago, the goal was always to get into the phone system. But with IBM, Watson and some of the voice tools out there, it wasn't a product we were comfortable to go to market with. So towards the middle of last year, we started rolling out an AI voice product and we have thousands of customers on the chat side, some of the biggest home service companies. We work with a lot of medical, dental, legal clients as well. Financial service, financial services, even a publicly traded company. Who? I can't say, unfortunately, but they're a Fortune 500 company. We have a big customer base. So when we rolled out our AI voice product, it was, I don't want to say it was easy to go to market, but we had a lot of customers who already trusted us and we were able to get market penetration pretty quickly and learn a lot very quickly as well. So we've done a lot to improve our onboarding. We've improved a lot of the conversation flows. And I don't know if you've heard any of our AI conversations or any of the listeners. If I know you share these on LinkedIn, I can always share some YouTube links of actual AI conversations. It's nearly indistinguishable from a human versus AI. And we can train the AI to sound exactly like one of our clients if they want to. And it's working out extremely well. And I'll pause there. I know I threw a lot of information at you and the viewers. I can open it up for questions, but trust me, I could keep going on some frequently asked questions that people ask as well.

Yeah, for sure. So basically the premise is when someone calls in and they typically talk to a csr, the CSR might fumble around, they might not be able to book the call. And also it's expensive to have a CSR team. And so what you're doing is you're having AI bots essentially that are answering the phone that have frequently asked questions, prompting them, answering. Is that kind of the premise behind it?

Yeah.

Yeah.

Your summary is extremely accurate. So I do want to clarify one thing with that. Our goal is not to take a bunch of jobs, it's not to replace full teams of CSRs. Because we've been on sales calls and you know, owners of these H Vac plumbing companies, they'll be like, oh man, I could just get rid of a couple of our CSRs. And we're always like, not the goal. That's not what we're trying to do. Our biggest use case is for after hours and overflow calls. So they don't want the phone ringing. For most of them are two or three rings. And as you know, Google, lsa, they even ding people if the phone is ringing for longer than 30 seconds. So there's a lot of benefits to that. With overflow and then after hours, some companies Use an after hours call center. Others, they'll have technicians and CSRs bring the phones home. We all know for a fact that does not work. So that's where we come in. And we'll even tell companies if they're thinking, oh, I could get rid of a CSR or two. We're always like, hey, you know, you can transition them to either be a dispatcher because it's hard to find good dispatchers in the space, or you could transition them to do outbound calls. We were talking about, I was in the office of one of our really big clients, George Brazil Plumbing and Electrical. They have a lady who does only outbound calls all day and we've seen the data she generates about a million dollars a year for the company, only from outbound calls. And she has a campaign for herself tracked in Service Titan. And we suggest them to transition some of their better CSRs to doing outbound. And yeah, it's a different revenue stream.

Yeah. So when, when a call comes in for people that haven't heard any of the calls, are they just like collecting data and then sending you over like a form that says like, hey, here's this person, here's their issue that they had, here's their phone number, contact them?

Basically, yeah, that's a great question. So there's a couple different scenarios when someone's calling in. Obviously, we know the, the biggest one is people looking to schedule an appointment. They're becoming a lead, and that's what we see most of the time. But there's a lot of calls where people are looking to cancel an appointment, reschedule an appointment. Maybe they're just trying to leave a message for someone. So when we're training the AI, we train it specifically for every single one of our clients. There's not necessarily an out of the box solution that people use. We have an onboarding form that it takes 15, 20 minutes to go through. But we're training for all of these different scenarios. And yeah, once a call is completed, it sends to the CSR team and then they can handle it from there. And feedback we've gotten from clients is it saves them a ton of time because the CSRs, they don't have to be on the phone for upwards of five minutes. Now, it's just as simple as a callback, confirming. Hey, I saw you were talking to our AI assistant and you wanted to book an appointment for this morning on your water heater job. So. So it's a pretty simple call and they just have to typically verify scheduling.

Yeah, that Makes sense. Is there any specific data that you've pulled from all the customers that y' all have? Like, okay, the average call should take about, you know, two minutes or something. And then the booking rate should be, you know, X, Y, Z. Or any interesting data that you could share?

There is some interesting data. So with a human taking a call, generally the calls are about four to five minutes. With the AI taking the call, what we see is the calls are about two to three minutes. So it's funny, when we train the AI, we're also training the conversation flow and the goal of the conversation. There's never any dead space or, like, dead talk time during a call, because the AI is trained to ask a question after anything that is said. So if someone's like, oh, the example call that I always use is some guy's always like, am I talking to AI? Or is this a person right now? And then the AI is like, oh, I am an AI assistant. All of our reps are busy at the moment, blah, blah, blah. But are you the homeowner of the house? So it's always following up with a question. Then he's like, yes, I am. So it's continuing that conversation. So there's not dead space and the call gets completed relatively quickly.

Yeah, that makes sense. What about, like, online scheduling? Is there a component of being able to do that?

Yeah, yeah. So we do have native integrations with a lot of the major CRMs.

Okay.

We push the data as an unbooked job, and that's what a lot of these online schedulers are actually doing. It's very tough to actually have it go into the schedule or the capacity board or dispatch board. However companies do it, there's a bunch of different ways, as you know. So we push it typically as an unbooked job, and. And then there is a little bit of human intervention where they have to confirm the booking and maybe they're following up with the customer. But we're setting that expectation on the AI call that they are getting scheduled, because that's the biggest thing. If someone's calling, they want to know that they're getting taken care of. And we set that expectation so they're not calling the next competitor and potentially turning into business for someone else?

Yeah, that makes sense. Is there any, like, buffer in the beginning that says, like, hey, this is AI Bot or something? So that way people. Because I've heard AI and it's sometimes it can be a little choppy or, like, a little bit, and then your senses get a little heightened and you're like, oh, is this really going to solve my problem?

That. That's a great question. So with our opening greeting, it's completely customizable. And we've been making some changes lately because at first it was, hey, thanks for calling XYZ Plumbing and H Vac. All of our team busy team members are busy at the moment, but I'm an AI assistant, blah, blah, blah. The second it says all of our team members are busy, they think they're leaving a voicemail. They think this is a voicemail recording. And about 30% of people were hanging up. So we were like, oh, man. After listening to hundreds, even thousands of calls, it's like, we should probably change this a little bit and see if we can improve these conversions a little bit. So a lot of times it's just, hey, thanks for calling XYZ Plumbing, H Vac. I'm an AI assistant to answer questions and book appointments for you. Let me know how I can help. So that intro message is pretty short now. And it's just setting the expectation that, you know, it's going to have a conversation with them. We don't necessarily have to say that it's AI. Some people like it, other people are like, yeah, just pretend like it's a normal conversation and someone's answering. So we tested it both. Both ways, yeah.

Have you seen that it matters on the call volume, you said it's like I said, a threshold. Or do you kind of just say like, hey, it's going to help with that regardless because you're missing calls, or do you have kind of a benchmark that you would say it makes sense?

I would say our ideal customer would be the bigger companies. If someone doesn't have more than like five technicians running, they're probably not going to see a ton of volume to really make the product worth the roi. I mean, at the end of the day, if they're booking one average size job, it's going to have a return on investment right away. But generally we work with the bigger companies just because they have so much volume. I mean, we have a couple clients who are getting upwards of 100 calls per day handled by AI. And we're not even in the busy season yet, which I'm expecting in the hundreds per day for some clients. And yeah, but I mean, there's other clients who are having two to three calls per day handled by AI. So it varies if, if you were.

Rolling over to like, you know, Ruby Receptionist or whatever those other ones are, they're basically going to be doing Something similar, but they don't know anything about your business and they're just going to be reading off a Frequently Asked Questions prompter, and they're not going to sound as compelling, potentially.

Yeah. And I'm not going to bash call centers because there's a reason people use them. And I know you're not bashing them either. Yeah, yeah. The thing with call centers, too, is even when you're using an outsourced call center, there's still a percentage of abandoned calls. Because the way a call center works is, let's say they work with a hundred different customers. I know the one you said is they work with more than that.

Yeah.

But they have a queue of all of these calls that are coming in. So they still have hold times. Even though it's going to an outsourced call center, it's still there's the opportunity for customers to be on hold for upwards of 30 seconds. I've seen data from customers who were using a call center before us, and there were customers on hold for four or five minutes before they even talked to anyone. So it's still a problem that exists. And we actually have a lot of call center relationships who are using our product to, you know, combat that problem. So it's something that is. Is a benefit to call centers. And, you know, there's some clients who also want that human plus AI element. So us plus a call center does make sense in a lot of scenarios, too.

Yeah, I could see that. Do y' all have anything about, like, rehashing or like the following up on estimates or unsold estimates, that kind of stuff? Have you got into that space yet?

That's a really good question. With some of our call center relationships, they are doing that. Generally, companies have a process for incoming leads and just, you know, incoming contact forms or chats, whatever it is. So they have a process for doing that. It's not something that we offer, but again, we do have partners who can help with that.

Okay, that makes sense. At one point, you know, Air AI. Have you heard of that?

Yes, yes, I do.

Yeah. I had built like something on that, like kind of a language model. And like, I was trying out some calls just to, like, my employees just like, hey, listen to this call. And they're like, yeah, a little choppy, a little weird. But, like, I could see how that would be so valuable. You have all these unsold estimates and someone's going to call you a, you know, a bot or something or AI.

Person for specifically doing outbound calls.

Yeah.

So that Is actually in our dev pipeline because people have requested it. And yes, the use case that you're saying with unsold estimates is probably one of the biggest ones. The other one is when technicians are on the way, it calls and says, hey, Bobby's on the way to your house. He should be there. And generally the average is 30 minutes. But you got a text message saying the exact time. So there's things that can be said. But yeah, we're building that out. It should be done. I hate to give estimates, but within the next couple months. I'm not even on the dev side of our team, but yeah, I got you our team of Del Mar.

I could see like, okay, the initial intake. That makes sense. The on the way calls, even like the next day calls, like, after the service is done, like, hey, how was the service? Just want to check in and see if anything was. Was everything was good or any issues. Same thing with the day before. Instead of just getting the text reply. See, it could be a nice call that says, like, are we all good?

Like, there's so many use cases, for sure. And as we both know, the happy calls that these companies are doing, it's so important because they've already closed the job, closed the revenue, but they also want to stay top of mind for the next time that they do need something done. So they're trying to do everything they can. Because I always tell our clients and I always tell other people in the space, we're in a commoditized industry. I'm in Phoenix, Arizona. I could pick up the phone and go on Google right Now, and there's 100 different plumbing and H vac companies that I could call. At the end of the day, who's answering the phone, who has the best marketing, and then who has the best customer experience to make customers want to come back to you? Because I've seen data of Companies doing between 20 to 50 million dollars a year. 60 to 70% of their business is from customers who have already used them in the past. Which is kind of crazy to think.

Yeah, it's crazy to think when service titan came out with the whole Uber thing, like, I'm on my way. You can track the technician and still people, when they see that, they're like, oh my gosh, this is amazing. Where has this feature been my whole life? And that really like what you're saying. It helps you stand out so much more from a lot of people.

Yeah, for sure. And it makes for companies that maybe have five technicians that are using something like that. It makes them seem a lot bigger than they are as well. So it's pretty cool to see how the trades industry has adapted to technology and really using it for. For the better.

Yeah. And had y' all always been in the trades, like with home services, even when you're back in the marketing agency world of leads and stuff?

Yeah, that's a great question. It's always been the core of our business. So right now we work with over 5,000 customers. I would say our biggest vertical is obviously home services, plumbing, H vac, roofing and about, again, 80% of the customers are from that. But the other 20%, we work with a lot of doctors, dentists, lawyers, financial services. So in the mortgage industry, we have some really big relationships, franchises in the home health space. We have one client who has over 300 locations in the home health space, and they have a very specific use case where we're on their careers page and they get hundreds of applicants every single month through the chat. And we do some intelligent lead routings for some of our franchises. So if they're chatting on the main company website, it's going to route to the nearest location. So the careers page is big for that. One of our mortgage clients who has about 250 branches, if someone's chatting on, you know, XYZmortgage.com it needs to route to the correct mortgage branch. So we have intelligent lead routing based on where the customer is locating to go to the correct location. So with some big franchise relationships, that's something that's really important.

Yeah, definitely. If you have a franchise all over the country and you have one website, then. Yeah, you got to make sure you go to the right spot.

Yep.

What would you say as far as, like, AI? I know we talked a little bit about some of the dev stuff that you have coming, but what do you see, like, in your crystal ball, if there was, like, down the line, is there any other, like, crazy things coming out or that you see with home services?

That's a good question. And I think there's going to be a lot more AI players in the space. The home service, you know, vertical is really good and companies are willing to adapt. And we always say people are pretty behind, but owners, marketers and GMs, they are willing to adapt. The one thing I will say that we've been focusing on the dev side is obviously the quality of the conversation. We're in a really good spot where it feels extremely human, like. But one thing that we were making changes to was something called, well, there's latency in the conversation. So when someone asks a question, the AI responds. There's a little bit of time in between where it needs to comprehend that data. It needs to go through some of our APIs and process it. And it took about one and a half seconds. So we would have a typing noise in between conversations. I know you've heard some of the calls, you've heard that typing noise. So as of a few days ago, we've kind of had a breakthrough on the dev side where we're now analyzing when someone is done with what they're saying. And there's like no latency. So it feels even more human, like, so we've really been working to make conversations sound as human, like as possible. And that's something that we're really excited about. But as for other developments in the space, I know like Surface has AI dispatching, where there's so many components to dispatching, where you want to know a technician's expertise, Are they good at selling water heater jobs, are they good at selling this type of job and then what's their drive time going to be? So there's so many components to dispatching and I know people are using service Titan's AI dispatching or Dispatch Pro, whatever they call it. So yeah, there's going to be some other advancements in the space and I know people are going to start to adapt.

Yeah. Cool. The other question that I had on here was about since y' all work with so many businesses, what have you seen aside from, you know, implementing some AI and automation stuff? What do you, what have you seen that it's been working really good, like from a lead flow perspective, obviously we mentioned Google local service ads, but do you have anything else that you're like, wow, this was really good that this company was doing this thing?

Yeah, that's a good question. So one of our growth models and the reason we work with so many, so many clients is because we grow a lot through partnerships. And our biggest type of partnership is with marketing agencies. So we have about 150 marketing agency partners. And when it comes to marketing, I think at the end of the day it is who knows what they're doing the best. And there's so many different ways to market. It's all about doing the little things right. Whether you're doing Google Ads, whether you're doing LSA SEO, I could probably count on one or two hands out of all of our partners of who is actually doing a good job and who really knows what they're doing. And we even tell our partners this. You know, you guys are, you know, your clients aren't getting a lot of chats and, you know, we're pretty upfront with them. We're obviously not going to talk bad about partners ever. But yeah, of course we really know who's doing a good job and who's not. So when people ask for referrals, we know who to send them to because we want them to obviously get more usage on the chat side and the voice side.

Yeah.

But yeah, at the end of the day, LSA is really good. It was a lot better in the past when less people were using it. But if you're doing the little things right, answering the phone within 30 seconds, you're making Google's algorithm happy. And then you can start to rank up on lsa. So at the end of the day, yeah, it's just all about doing the little things right. Cool.

Yeah. One question that just came up was you mentioned about chats and getting traffic to your website. Do you have a conversion rate that you typically recommend? Like, if you get 100 people to your website, you should convert chats or something at a certain percentage or. Have you seen any numbers like that?

Yeah, that's a really good question and definitely something we've heard in terms of like pure web traffic going to your site and actually converting into a chat, that's tough to say because the quality of traffic is so important. If 100% of that traffic was from Google Ads, conversion rates are going to be super high. But if you have a ton of traffic from blogs or whatever it is. Yeah, you know, we're both marketers, we know conversion rate is going to be significantly lower. But the data that is important when someone starts a chat and they go through and actually finish it, because the responses are so human, like, and they're so instantaneous, versus a live chat solution. Our chat completion rates are about 97 to 98%. And when someone's like, well, what's a completed chat? We're capturing what the customer needed. We're capturing their name, their phone number, their email address. So we're capturing all of the lead information and we have data to show. So it's easy for us to talk on these numbers. Cause we can pull up any account that's getting a ton of volume and show them, hey, you know, they had 982 chats completed out of a thousand. So.

Wow. Yeah, I think that was always the hesitation is like someone comes to your website, they start chatting in and they leave your site and then you lose the chat and you lose the lead.

Oh, a hundred percent. And that's why AI and the instant responses is so important because we live in a day of day and age where everyone wants instant responses. They want instant answers. No one wants to wait for 30 seconds. I mean, we'll be scrolling through TikTok or Instagram and see a video we don't like. We don't like it within three seconds onto the next video, onto the next video. So we have these super short attention spans and it's the same thing when you need something urgent done in your house.

Yeah. So if someone wants to hear what this sounds like, they can just go to the website. Zero Talk.

Yeah.

Dot com.

They can go to our website.

Yeah, there's options on there, right?

Scroll down a little bit. You can hear a bunch of real recordings. And these are real calls handled by AI. We've bleeped out, you know, some of the, you know, identifiable data with addresses and the company names. But yeah, they can listen to real calls.

Nice. Do y' all have any like free trials or anything? Or people go onboarding, like what does this all look like if someone's interested.

Yeah, yeah, great question. We don't do free trials because there's hard costs on our end. But we are month to month. So we're not going to lock people into year long contracts. They want to try it out for a month, we can do that. For partners, we waive our onboarding fees. So if someone comes through a partner, we have a thousand dollar one time onboarding fee. We will completely waive that because generally our partners are getting us accounts at scale so it makes sense for us to waive that. And then the way our pricing works on the AI voice side, it's $500 a month and then a dollar per AI minute. And if a company's a certain size, we can do things like throw in 100 minutes. Something interesting we've been doing lately. If there's any pushback on price, and I'm saying this now, anyone listening? We'll throw in 100 minutes if you're listening. So something we've been doing is they're like, oh, if I don't use my minutes, can those roll over? Generally the answer is no. But as we go into the summer months, we will roll over minutes up until August because we know summer is going to be super busy and we know those minutes are going to be used.

Yeah, okay. I mean that, that sounds good. Like if someone's like, hey, I don't have someone answering my phone after hours. Even if I can get, you know, a couple jobs out of this, like, why not try it for a month or two, even if it's just in the summertime, and then start to play it out and see how it would go. Nice, man. Well, yeah, you know, we know a really good marketing agency called Flash Consulting.

So we need to get on this.

We need to get this agency so we can get hook our people up.

So, yeah, totally.

Cool. Yeah, we'll talk. We'll talk after this. Awesome. So it sounds like go to zero talk.com z y r a talk.com and then they can sign up or at least listen to the stuff and see if it's the right fit for them.

Yeah, absolutely. And if they want a demo and they want to talk to one of our sales reps, or if they want to talk directly to me, I'm one of the founders. You know, fill out the little thing on top, put in your email or I think you can scroll down somewhere on the site and put more information. But we will reach out to you and we'll book a demo. I've heard this is kind of crazy because I know there's, there's like one other player in the space, in the plumbing and H vac space. The thing is, they are pretty behind. I'm not going to talk bad about them because I know they're getting started and other people have mentioned them. They don't have the back end infrastructure built. There's a lot of nitty gritty things that need to be done with notifying technicians after hours. We've built so many different things specific to home services, but other people in the space, they're smaller and they're like, yeah, I couldn't even get a demo with them. So anyone who wants a demo, we have a sales team. We're not like super pushy with our sales, but we'll go through a demo, talk about your business needs and all of that.

Cool. Well, thanks, Brad. This has been great. It's so good to get to know you more and learn about the product and stuff. So thanks so much for joining us.

Yeah, I appreciate it and excited for people to watch this and give us some feedback.

All right, thanks, man. Talk to you later.

Yes, you too.

Bye.