Kevin DeMeritt - The hidden requirement for conversational AI at scale

Rachel Thorne: This is Dialed In with Derek Simmons and Rachel Thorn. Let's talk compliance. Derek Simmons: OK, so welcome back to Dialed In. Today is fun. We've got Kevin DeMeritt in the house talking voice AI. Speaker 3: Yeah, this one's stacked. We're going from his background of voice AI and how we even got into this space to that messy speed to lead follow-up. Derek Simmons: Exactly. Then we'll dig into why compliance has to be the first priority, how TCPA risk actually shows up on your balance sheet, you know what I mean? Speaker 3: And we'll hit trust and voice AI, plus what it looks like when... When compliance is baked into the architecture instead of bolted on at the end. Derek Simmons: So anyway, let's jump in with Kevin's origin story and how he first got pulled into voice AI. We want to hear from you. Submit questions via the web form in the description or give us a call at seven four seven nine four six seven six zero two and leave your question. Don't be shy. Our AI assistant makes it super easy. Speaker 4: To start, what first drew you to this space? Speaker 5: Well, I was running a precious metals company that I owned for about 35 years and we were drawing in about 8,000 to 12,000 leads a month from form fills online and our salespeople, which we had about 125 salespeople at the time, were just getting behind in speed delete and then the follow up. So I started looking into voice AI with another company about five years ago and decided that I thought I could do it a little bit better with TCPA compliance and the security, getting the voice to sound the way that I wanted it to sound and really be a sales and marketing help as opposed to just something built by coders. So we started that about four years ago. It had a couple of years of development, and then we launched about a year and a half ago with 2X. So really proud of what we built. Derek Simmons: That makes sense. When you were building it, what had to come first? Compliance, security, or the sales experience? Speaker 5: Well, compliance had to come first. Compliance, security, you know, because that gives you a lot of legal exposure. So that's always the first conversation we hear from almost anyone out. out there because it's not just software sitting in the background. You are making live phone calls, you're handling customer conversations, and you're touching real customer data. So if you're not compliant with TCPA, the Telephone Consumer Protection Act, and if you're not managing opt-outs correctly, if you're not controlling when that call happens at specific times that are regulated by states and the Fed, then your company's exposed. And if you're not storing transcriptions with the outbound call number and the numerous other required data points that are mandated for five years, the hold period, you know, you're just out of compliance. So at that point, it's no longer a technology issue. It becomes a business risk and a $500 to $1,500 fines per call. It could put a lot of businesses out of business. So you kind of have to start with security and compliance first. Derek Simmons: That makes the stakes clear, Kevin. What does trust actually mean in voice AI for you? Speaker 5: You know, most companies focus on really to me how human the AI sounds before they ask whether the system is actually safe to deploy at scale. And the real execution question is simple, in my opinion. Can the business trust this platform first and then what are all the cool bells and whistles second, right? The voice and some texting and all those kinds of things. And so quite frankly, most voices to me are starting to sound the same. It's starting to become a commodity. The millions spent to make the platform compliant, That's the rarity. And if customers do not trust it or the regulators do not trust it or leadership doesn't trust it, none of the other metrics really matter. So trust to me is predictability. How the system behaves correctly each time, does it follow the compliance rules every time, it respects the day of time restrictions, the do not call list, the opt-outs, the data handling requirements, that's the foundation. So at 2X, we built compliance into the architecture from the beginning, the TCPA compliance, the HIPAA compliance, the SOC 2, Type 2 security, the real-time spam remediation, because if you're scaling conversations, you cannot leave compliance to risk. risk. Derek Simmons: That's helpful. Once trust is in place, what becomes the next priority? Speaker 5: You know, I think that the operational usability, can the business actually run the system at scale reliably? I think that's the next big. hurdle that a lot of companies start bumping into because there's a difference between sending you a demo and it's sounding good and then having it make hundreds of thousands of phone calls. And in one month, we made 800,000 plus phone calls for Bosley Hair. Your system has to perform at scale if you want to be able to do something like that. Derek Simmons: What usually breaks first when a company goes from demo mode to real scale? Speaker 5: I, well, compliance usually breaks down because you need to do that at scale. Every call that goes out, we need to make sure it's going out at the right time. If you're putting it through the do not call list, then you have to run through that in real time. The voice has to be able to or your server has to be able to. make sure that it can make 800 a million calls per company you know each and every month and so usually the systems start to fail and if they start failing for one company they usually start failing across the board so you know you could be a smaller company and have an architecture with some large company on there that's eating up a lot of the time um making phone calls leaving voice messages sending texts things like that uh and even if you're on a different server the main server that's running all of it if it breaks down then you break down because somebody else has made a lot of phone calls so you you have to have the architecture and the backup and the load balance all correct to get this Rachel Thorne: system to make sure that it does what it's supposed to do. Derek Simmons: What does the right architecture look like for that? Rachel Thorne: Well, first, you need extremely large servers and you need the load balance in different... parts of the country so that if one area goes down that your customer all of your customers don't go down that you can load balance it over to some some other location and usually need somebody like Amazon or Google Cloud to make sure that you can Put all of that together and have it scale correctly over time. I mean, you continue to add people if you're a successful business. And so that architecture has to be there to make sure that you can scale with it, not only from a compliance and a security standpoint, but also from a performance and operational standpoint. Derek Simmons: That sounds like the foundation. Where do companies usually underestimate the infrastructure work most? Rachel Thorne: Just how complicated it is to. So get your training for the AI to do what you would like it to do consistently and at scale would be number one. A big hurdle is getting the TCPA to be as compliant as humanly possible and following all of the rules from the state level and the Fed level and trying to keep up with all of that so that the platform is compliant. That's a big hurdle and making sure that security over and over and over and audited continues to be there so that your data coming into the 2X system or any system out there really is protected. You just don't want your client's information out, obviously, on some privacy security situation where there was a break-in on the system. I think the training of the AI, making sure that it conforms to the way your business runs and your knowledge base and your objections and closing and everything that goes into that is a part of that training and then compliance and security after that. Derek Simmons: What should leaders get right first when they deploy voice AI? Rachel Thorne: Well, it's going to sound like a broken record, but it's compliance and security. You have to get that right first, so they have to make sure that, you know, that's built into, you know, the platform. After that, it's making sure that your CRM and your tools are talking in real time back and forth to you. Either 2X solutions or whatever platform you're working with, making sure that those tools are working so that there's real time. I mean, a perfect example would be, you know, we're talking to a customer and they ask us to opt them out. We need to opt them out immediately. That's the rule. So we would opt them out and then send that information back to the CRM immediately after that phone call so that they could opt them out so that one of their representatives doesn't accidentally call and you have a regulatory concern or some class action attorney coming after you. So I think that's what you have to get right first is just making sure that the architecture is correct for your system. And that's just basically APIs and webhooks is making sure that some technology person is there to kind of, you know, hook it up and run it. It's not that difficult, but somebody has to be able to do that and make sure that it's correct. Derek Simmons: What do teams usually miss when they connect the AI to their CRM? Rachel Thorne: I don't know if they really miss too much, quite frankly, Jordan. You know, once. So we're very white glove. So maybe it gets missed if it's a DIY, you know, voice agent that a lot of people try to plug in and do themselves. But for us, we do a lot of testing. So there just isn't a lot missed, I think, when. You have the team there making sure that you're integrated correctly or we jump in and help out with that integration. It's not, like I said, it's not complicated. You just need to make sure that it's done correctly and make sure that there's a little bit of testing done so that information is flowing back and forth and you can see it and confirm it. Derek Simmons: What is the one thing leaders should remember before deploying voice AI? Rachel Thorne: You know, that the bottom line with voice AI, in my opinion, is. We are replacing the repetitive human tasks that should be done by AI, which can do it better, faster, and more consistently. And if you look at it that way, then there's an ROI sitting in two different places. One, the efficiency of getting a lead in, getting speed to lead and a minute or two after that lead is submitted, how many more leads are you going to contact, which turns into real business at the end, which is your ROI? And then how many people can you redeploy from repetitive tasks? over to high-level thinking tasks or closing tasks to make sure that you are generating that revenue. So I think that's what we like to sit down with businesses and make sure that they understand there's two parts of this that we're going to try to help solve for you, and then your business can scale much faster with systems that are repetitive done by AI, voice AI, the voice AI calling. and leave the higher level tasks of closing and decision making and answering high level questions to individuals because people waste a lot of time leaving voice messages and trying to dial people and you know 80% of the time you're leaving a voice message a complete waste of time for people but getting that person on the phone maybe qualifying them a little bit or creating an appointment it's just a complete waste of time for people when you have systems like this and so we want to Speaker 6: make Rachel Thorne: Make sure that, you know, leaders understand that and what are our KPIs so that we're aligned and making sure that the system works and getting an ROI that they require. Derek Simmons: Kevin, what is the one takeaway leaders should remember? Rachel Thorne: You know, I think that when we see most of the companies that we work with, they don't actually have a lead problem. They have a conversational handling problem. You know, we've seen companies re-engage, you know, 10 and five-year-old dormant leads and generate over $200,000 in additional revenue from databases that they thought were effectively sitting idle. I mean, nobody was going to call that. We've also seen AI-driven workflows generate $60,000 and $80,000 per month in additional profitability simply through better operational. consistency, the follow-up, the speed, the lead, and all the things that I think sales managers, CEOs, and marketing people see on a daily basis. You know, people create relationships and systems create operational consistency. I think What AI does best is free teams from, like I said, this repetitive communication work so they can focus on high value conversations. And that's where scalability really starts to happen. So we turn your repetitive phone tasks into a predictable, scalable revenue system. So in my opinion, tomorrow's winning companies will realize they need to stop wasting human talent on these low level phone tasks that AI does. faster better more consistently so that they can free up their human talent for these high-level decisions and closing more business Derek Simmons: Totally. Derek, Kevin really flipped the usual priority list on its head. For him, number one was compliance, number two was security, and only then do you earn the right to choose sexy features. Speaker 3: Yeah, and he was blunt about it. If you don't design for compliance from day one, it's exactly what breaks the moment you move from demo to remote to real value Derek Simmons: Exactly. His point was, in a pilot you can babysit the system; at scale you need guardrails that work when no one's watching. Speaker 3: I also liked how he defined trust in voice AI. Not just it works, but it behaves the same way at 10 calls or 10,000, and it never surprises your legal team. Derek Simmons: So, for leaders, his implied checklist was one, lockdown compliance. This to prove reliability at scale; three, then optimize experience. Speaker 3: And don't underestimate the plumbing. Telephony, logging, monitoring, that's where people get burned. Derek Simmons: Stay with us. We'll wrap this up in the outro next. So if Kevin's story of building voice AI from that lead follow-up headache to compliance first tech hit home, remember this, trust starts with consent and proof, not hype. Speaker 3: Right. And if this opens your eyes to those TCPA and business risk apps, subscribe, drop a quick review and share this with your legal team. Derek Simmons: Yeah, and hit 2xsolutions.ai to see how compliance first First Calling Works. Thanks for listening. Speaker 3: See you next time.