Sponsored by the Shift Group – Shift Group is turning athletes into sales professionals. Is your company looking to hire driven, competitive former athletes? Shift Group not only offers a large pool of diverse sales candidates from entry level to leadership – they help early stage companies in developing their hiring strategy, interview process and build strong sales cultures that attract the best talent for early stage companies.
Sponsored by the 4-Step Guide to Delivering Extraordinary Software Demos that Win Deals – Click here and because we had such good response we have opened it up to make the eBook and Audiobook more accessible by offering it all for only 5$
Sponsored by Diabolical Coffee. Devilishly good coffee and diabolically awesome clothing
Does your startup need strategic technical content? The team at GTM Delta delivers SEO-optimized, compelling content that connects your company with technical users to help grow your credibility, and your pipeline.
Brian Podolak is CEO and co-founder of Vocodia and an experienced and visionary contact center innovator with a background in electronic and mechanical engineering.
We chat about the Conversational AI revolution, what’s next for the call centers ecosystem, how conversatioanl AI is growing and disrupting traditional sales teams, and how to leverage AI for business in an authentic way to improve the human experience.
Check out Vocodia here: https://vocodia.com/
Connect with Brian via email to brian at vocodia.com
Thank you for a fun and informative discussion, Brian!
Transcript powered by HappyScribe
Want to advertise your business in a cost effective way? It’s time to give podcast advertising a try. Research shows a high rate of podcast listeners made a purchase as a result of an ad they heard on a podcast. Visit podbean.com/brands to launch a cost-effective podcast advertising campaign in minutes that’s P-O-D-B-E-A-N.com/brands.
What’s happening? Welcome back, listeners, crew, family, friends, and newcomers to the podcast. Holy Moly. We are rocking out the chart, so thank you so much for listening. Watching. Make sure you also head over to the video version of the podcast so you can go to youtube.com/discopossepodcast and you can see the whole conversation as it happened, which is super cool. All right, this is Brian Podolak from Vocodia. Brian is really, beyond being a fantastic human, has an amazing radio voice. He’s just somebody who’s solving a really neat problem, has a background in engineering, but Vocodia is doing stuff around conversational AI. And this is not the, you ask it a question and it pops back a thing, where you’re calling into a call center. This is actually the reverse, where they’re having outbound conversational calls, that pre-process things to help drive better conversations and really, ultimately increase the flow of success for customer experience. Super cool. We dig into the tech, we dig into his own background, how they solve the problem, and just their approach as a company. Really neat story. Definitely a must listen. So I hope you enjoy it. And by the way, these kind of amazing conversations, which I’m about to have a great announcement about.
Super cool. So hang tight. That’s coming up in a couple of weeks. But in the meantime, talking about great things, the folks that make this podcast happen, that support us, I got to give a shout out to our fine friends who sponsor the podcast over at Veeam Software. If you want to get everything you need, whether it’s on Prem, in the cloud, everything you need for your data protection needs, this is the place to go. Go to vee.am/discoposse. How easy is that? Seriously, just go to vee.am/discoposse. Check it out. They got lots of great tools, whether it’s SaaS protection for stuff like Microsoft Teams and Office 365 and much more that’s coming down the pike. Congratulations to Veeam on lots of really great stuff. They’ve been doing great new products and capabilities. And another thing you got to make sure you do is protect your data in flight, which means wherever you are, make sure you’re running a VPN. I know I use one. And this is not the I don’t need to protect free speech because I don’t say bad things. This is literally your identity is flying back and forth. You’re putting informed data.
You’re filling stuff. You are at risk. Use a VPN. It’s just a great practice. I use ExpressVPN. I recommend you do the same. You can go to tryexpressvpn.com/discoposse and you can join me in the battle for protecting your privacy. All right, let’s get to the good stuff. This is Brian Podolak.
Hi, I’m Brian Podolak, CEO and-co founder of Vocodia here on the DiscoPosse podcast.
You got a perfect radio voice, Brian. This is it. You make it easy for me. I just get it going. I’ll come back in an hour. You’ll have a great show all on your own.
No, I like to talk, but I like the conversation for sure.
Yeah, this will be good. Thank you very much. This is a real honor to share mic with you. I really dig what you and the team are doing. And I’ve had a few folks on that are sort of in the area of this idea of conversational AI, but it traditionally is at the drift chat type of layer. And being able to go farther and really see the power of what we can do in voice technology is super exciting to me and a bunch of people that I’ve talked to. So we got a lot of people who are going to be sitting in. And the good thing is we’re on camera. So it’s not like you sense a voice assistant to do this for you, Brian. But the next time we should actually bring a third party, have some Vocodia technology helping us out with the podcast.
100%. Every time I do a demo, people ask, is that really you doing the demo? Yeah, all the time.
Well, thanks for having me, Eric. Appreciate it.
Yeah. If you want to give a quick sort of bio and your background, Brian, for folks that are new to you. And then we’ll jump into the Vocodia story.
Absolutely. So make a long story short. About 17 years ago, I was relocated to Costa Rica to help manage a call center a friend of mine was running. And make a long story short, within a year, he decided, I don’t want it anymore. You take over. And we went from about 20 agents when I was there to about 600 at our peak. And I always said, I love the business, but if only I could do it with less people. Right. And the challenge with that call center business is it’s easy to find five or ten people to do a good job. A lot of times it’s harder to find 150 or 200. And so between things of turnover rate training, just hiring, HR, et cetera, et cetera, we’ve always thought this would be a neat concept. And the genesis of this was that we had a client that we had what we called a retention campaign. Basically, people who had a recurring charge on their credit card would call us to cancel. And our job was to try to have them stay on just one more cycle. Make a long story short, every call you get in, most people are very upset.
Nobody says, hey, you’ve been charging me $40 a month. Can we work on it? It’s usually I hate you. I want to kill you. These kind of things. So it’s a very hard campaign. Our turnover rate was approximately 900%. What we did is we had the agents as an idea to start their first line of Hi. My name is Brian. How are you doing today? Each agent pre-recorded it and press the key on their keyboard to say their opening line. Well, the agents loved it. Immediately they had like ten or 15 of their most common lines pre-recorded on the keyboard. So now you took a job where you’re getting beat up all day and made it a little bit fun and just took off a little bit off of you. And having that energy on that initial call, as you can imagine in the morning, compared after six to 7 hours, getting beat up, being pre recorded made a noticeable difference. That was kind of the genesis, I like to say, of where we are today with our technology. So my background was running large call centers for very big clients, consulting with other call centers, literally worldwide.
We owned our own centers in Costa Rica, Panama, the Dominican, et cetera, a consultant from the Philippines, India, and a lot of other countries you probably never heard of, like Ghana, et cetera, that we actually had centers into. And it’s a very, very great business. It’s a very big business still to this day. And to be able to do that internationally was a lot of fun. So I’m very happy on where we are on the journey today.
Now, this is an interesting thing. Boy, could I use that. I could do my podcast that way. Just have everyone just be like, hit F1, you know, that’s a great question. How does that affect your business exactly?
Tell me your history. Right.
The story of it and the reason why it’s been meaningful to customers and to you know, your folks that use the platform is super important. The raw technology is incredible. Like the excitement I get, I’m like a nerd. So I get a nerd. And I’m like, oh, all right, we’re going to talk about some real nerd bits. So the good thing is we can kind of cover both bases. But really the truth is, for folks that most folks that are listening, we can nerd out all we want. But in the end, the story of how it actually has delivered meaningful value to a human is the best part of technology. And when people hear stuff like voice assistant and AI and conversational voice centers, they worry. The first thing is always like, oh yeah, you’re taking away jobs. And so we’re going to get right to the good stuff.
Let’s do it.
Let’s dispel that rumor.
Well, Ironically, what you’re talking about is true. When Jimmy and I first started having a little bit of our breakthroughs and I’ll nerd out later with you. But we started having a breakthroughs was pre Covid. When we first got a working model was right before COVID was hitting. I was actually a consultant in the Dominican at the time and had to come back because Covid was just starting to become an issue where it was sort of in the early days. I remember saying, I’m not going to get back home if I jumped on that plane, got home, and people were just starting to wear masks at the airport. And so we got over here, the narrative when people heard about the technology and started hearing the demos, they said, oh, you’re going to replace jobs, you’re going to take jobs from people. Well, now fast forward from Covid, and you have what a lot of people call people who don’t want to work. And where we’re focused on is not replacing people at all. It’s augmenting existing processes. It’s helping fill the holes in an organization. So we definitely are not there to replace anybody. We’re just trying to help fill in.
And more importantly, because it is so hard to find people now, we found a lot of our clients and prospects we’re talking to, are actually settling for mediocrity. So, for example, let’s say you have 100 people in the phone and there’s 30. You’d like to fire them today and get rid of them, and they’re just not working out. But your fear is you can’t replace them, so you keep them on the phone. So what happens to your client now? The end user is now dealing with somebody who’s probably not doing the best job for you. So everything on our design, not technical at all, was thought about the end customer experience first and making sure that experience was perfect. A lot of people get compare technology to IVRS as they’re trying to understand it. And I always like to say, I did my dad test. When my dad was able to get that call and go all the way through and had no problems. That’s when we knew that we had a working product. But we’re definitely 100% right now. We’re having more and more clients come and say, listen, I need 20 more positions, 50 more positions, whatever it is.
And that’s been our strength right now.
Yeah. I think the thing we have to remember is what is the actual purpose of that call center experience? It’s a way that you can route to somebody in IVR have been the death of any hope for that for so long because it’s always like, how many times do you get on? And you and I, we’ve been around the industry for a while, Sue’s like, hello and welcome to Ex Call Center. I’m hitting the zero button downtown. I’m doing whatever I can to exit out of the menu. I’m trying to trigger it to fail. So, like, one moment we’ll get you a person. Yeah. I want to fast forward to the human part because they’re true-false, if-then-else, they aren’t guided conversational processes. And this is the gap that people don’t see. So we do this upfront kind of like pre interview through the IVR, which just enrages the person to the point. Then by the time they get to the human, they’re like, look..
You bring up a good point, Eric, and that’s exactly what we try to avoid. For example, about two months ago, I had an airline issue. I couldn’t change my ticket and I couldn’t do something online. It was one of those errors, how to call in. And this particular airline, when I called in, it was a three-hour hold time. They didn’t even have that we-can-call-you-back option. So I’m on hold for an hour. I’m going around my damn hearing this music. I was getting so frustrated. Finally, somebody picked up on an overseas call center. And unfortunately, due to this COVID thing, we have a lot of people who are not working at home. So you’re in a third world country, not even in a proper call center, trying to work from home. And this poor person, let’s face it, they’re trying to make a living, provide for their family. Right? And I’m talking to them and all I hear is this and I’m trying to talk to them and get my conversation out. And now I don’t want to hang up and start over, for the love of God and start the process over.
So my experience was horrendous. Again, the end-user experience has got to be as good, if not better than a human for a new technology to be adapted. People don’t want to adapt a new technology if it’s lesser quality. Ironically, they’ll settle for lesser quality people before they’ll settle for lesser quality technology. So we had to make sure at any time the AI got stuck. And AI in general has to be not to geek out too much, but it has to be trained. Just like when you hired somebody, you have to train them, they have to learn. So if the AI does get stuck or it gets asked a question, it doesn’t know how to reply to, one of our preferred options. And we have many. One of the preferred options is, hey, you know what, Eric? That’s a great question. I don’t have the right answer. Let me get you a supervisor and immediately get them to a human. So it asks a question one time if it doesn’t know or if it’s stuck, it immediately transfers. So we first do an implementation. There might be a little more transfers in the beginning as the AI learns.
But one of the ways that we speed that up and again, to get immediate better customer experience, because again, nobody wants to implement an AI and have it learn with real customers, is we just upload 500, maybe 1000 recordings of agents on the phone to help train the models. So we basically take three to six months of training. And we can do it in a matter of days now through our AI modeling process. And that makes that initial rollout to give that end user customer a much better experience than any IVR tech technology that’s out there.
Yes. And this is really the differentiator that we’ve got with the technology available to us now. It’s much more both democratized as far as availability and then commoditized in that it’s being broadly used. So then the pricing is no longer this real sort of platinum pricing for a coal experience. You’re able to now leverage back end technologies that are continuing to evolve now even beyond before, where you’d be like Brian and his team are developing their own machine learning. They bought a data center, they filled it with stuff, they’re teaching it. You can leverage other technologies that have broad access to scalable architectures. And then that lets you be your business now can empower the person who God bless that poor call center person or a call center manager or even a CIO for a company that has a significant call center presence, they shouldn’t have to care about the technology that could do something for them. They should be buying services.
Correct. And one of the things that we thought would be our key to we had a couple of challenges. In the beginning is, especially if you’re doing an outbound presales call. So let’s just say you filled out a lead to buy a product or service. You actually want someone to call you back and actually speak to them about that product or service. So when you get that call back, you want to be assisted and make sure you get routed or have the right information sent to you. So for example, you’re looking for a quote on health insurance. An agent is going to call back, verify if you’re under 65 or under 65 to see if you need a Medicare package or one of the Obamacare packages. And then there might be some budget restraints. So there’s a lot of questions we could have a true conversation. And what’s the great thing about it is this? When we first started doing this, we have latency issues. Now if you do an inbound call to a bank and you lost your credit card and they say, okay, let me pull up your information on the system. We’re used to delays so you don’t notice a latency in a conversation on an inbound customer service that call at all or very little but outbound sales.
If I say, hey, how are you doing today? And you say, I’m fine, and it takes me 4 seconds to reply to that, it just kills the conversation. So on the technology side, we developed our own proprietary MLP, which is the brain of an AI. And it’s the only AI we know that’s been 100% trained only with call center sales conversations. So that was all of the training we gave to it. Then the secondary thing was getting that latency down. So now we’re literally about the 20th of a blink of an eye or about five milliseconds per response. But then we connected to a VoIP switch. Well, now VoIP has its own set. So we’re very fortunate. We were able to get Omar Khan, our CTO, who came on and said, listen, I’m not going to work for a start up. I hate start-ups. I’ll be a consultant with you. And four months later, I wore him down. He’s now our CEO, and he developed a voice switch optimized for AI, not just our technology, but AI in general. And our response rate on the phone calls is just extremely quick. We don’t have I don’t want to say no latency, but it’s basically almost immersed.
You cannot even hear it. In fact, the last recording I sent out, Eric, somebody thought I edited it. It was that fast. So we actually do live all the show. Yeah. But as you said, go back to your other point. Now you have the customer shouldn’t be worrying about latency, shouldn’t be caring about VoIP switches. They want to implement this and get their customers sorted and helped. One of the other great things about the AI is it doesn’t want to hang out and be your friend. So, for example, Eric, you call in, you just like, I just need a health insurance quote. I’m going to use health insurance a lot because we just got out of open enrollment. So you’re going to call in and I go, Eric, where are you from? And you say, I’m from Jackson, whole Wyoming. I go, oh, my gosh, I have a brother and wax. And I start talking about a restaurant down the street. And you’re probably being polite, but you want to get to the point, right? So what happens is one of the stats that people look at in a call center is, what’s the average time it takes for an agent to handle a call.
And let’s just say you have an average of, let’s say five minutes. So many times when we implement that number goes down by 30% because we’re staying on script. We’re getting to the point. And you as a customer, getting what you want quickly. You’re not hanging out. You’re not having a blah, blah, blah. You’re getting to the point. Customer experience is happier. We’re spending less time on the phone, getting more transfers to wherever the sales reps are. And the great thing about it is it is a turnkey solution. We’re not a framework. So you literally just give us data, we click some buttons, and a week later you can have AI up and running. It doesn’t require a forklift upgrade into your team. It doesn’t require a whole bunch of any hardware. It’s literally give us a phone number to send calls to. We handle the rest. So in answering your questions and jumping around a bit. It’s kind of where we are today. The first true turnkey implementation of AI that people can get up financially risk free because we don’t let our clients pay us a Penny until it’s implemented and turned on.
If they’re not happy with the results or it doesn’t meet those expectations, they don’t know it’s a dime. And that’s how we’re able to get our foot in the door. As a young start up, we’re very confident in our skills. Of course, there’s a lot of understanding of clients scope of work and requirements before we do that. But we’ve only lost one client thus far since we started.
Ask me why I lost them.
Exactly. Now, that’s fun part. Now when you say you’ve lost one, that you’ve already laid the groundwork that I got to ask now. Yeah.
So the one client we lost was that our AI was getting them so many transfers. And as a startup and like every company, you have your requirements of what you need to do business, our requirements. We require you to think of ten physical agents on the phone. So ten licenses of our software is basically like having ten people on the phone, virtual agents, if you will. Well, those ten agent minimums were making so many transfers go into their call center, they couldn’t handle them all, so they had to pause us until they could ramp up more to be able to handle the amount of calls we were sending them.
It’s an Eli gold rat problem of throughput and the theory of constraints that was highlighted. So that was just purely market timing with that customer, which is really 100% that. Yeah, it’s the right reason actually stepped back, too, is as a business owner, you have to have the respect to be able to share that story and recognize where it wasn’t the fit. Inevitably, once they’re ready, who do you think they’re going to come back to? Right. They’re not going to try and ramp up this front line again. They’re going to say, let me get my tier two team really ramped and ready to go. And then let’s put front line and have it be that better experience because it’s like when you talk about call duration. So I’ve worked in insurance companies for a long time in technology, but became very familiar with the business and the call center business side of it was really amazing to watch because it’s a tight ship. They’re managed differently. It was a little hard to watch sometimes because the metrics that they’ve got to get, they have heroic numbers that they have to hit consistently. And like you said, you get these conversations.
The point where going into a chat where someone is from and getting them happily engaged is tier three type of level. And that should happen because tier one should be like, I just want to get to my answer. Right. Tier two is either they got to the answer and they need a little more detail or they’re really irked and they need a human. So then you’re handling them. Tier three is okay. I’m going to need my manager. And we all know that I’m going to have to get my manager on the line. They’re now like hostage negotiators. That’s not a call center anymore.
No, you’re right.
That’s like Chris boss suddenly. What do you think that I can do about that? You’re bringing the guy on that’s got the late night DJ voice. We’re going to talk about your problems.
Good old Johnny Fever right from back in the day. Yeah. And our technology does great on the tier one. Tier two, tier three. I don’t think we’re there today. And of course, we always look at competitors and try to learn from everybody. And you have a lot of conversations. But I think that there’s a point where people do know they’re talking to a bot or an AI system, and there’s nothing more comforting than talking to a human person. Right. As much as AI is great and can help us and definitely augment our processes. And for example, if I’m a customer and I just want to say I’m checking on a tracking number, I’m checking on a refund status, I got a credit card loss, these kinds of things. Whether I’m speaking to an AI in person, I care about, did it get done? Did it get an effective and at the end of the day, I feel good. After I hung up that phone call, hey, you know what? I lost my card. It’s a horrible situation. But I called and hung up. I actually just replaced my debit card two days ago, and I called in and guess what I had to do?
I had to type in my number, pushing the buttons on the phone, had to push in this numbers. And it was such a frustrating experience. And again, you already have a lost card. It’s already a little frustrating. And now you have to go through a process and it’s not thought of, hey, how do we support customer just went through something. It’s bottom line from a bank is how do we not have to speak to this person and make them go through the process and hit the numbers on your phone, which is called the DTMs? It’s not a happy experience. And especially if you miss Typ or hit two at a time, you got to start over. It’s just not exciting. And then you hit it in and have it repeated to you. So there’s so many little things in a customer service or sales organization that not only is it our AI to try to make that process better, it’s the experience that I have and other people in our company of working for centers for 15 or 20 years that’s making this technology work. So it’s funny, in the beginning, people are like, well, you’re a tech company that happens to do.
And that’s a big thing in tech companies, right? Your tech company that does sales, your tech company that does health insurance, or a tech company that does this, we started off at that way. I said, you know what? No, we really are sales first. We’re using the technology there. But every piece of technology we’ve developed today in our product roadmap over the last twelve months is all about helping a sales process. Now, after I do a sale, I want to keep that customer. So customer service is a piece. But it all started with that sale to a client. You used to use my service, Eric, but now you don’t anymore. Well, why not? Let me reach out to you six months later, Eric, can I do a quick survey with you? Can I get you to come back? Right. And a lot of times we’re talking to clients like, man, I want to do that, but it costs too much. But now with AI, we have a service that maybe we weren’t even doing before that makes that customer happier. Wouldn’t you be happy if somebody that business that you’re working with actually called and said, hey, here’s our product roadmap.
What other robots do you want to see? What’s the feature you want to see what’s the KPI we should be doing a better job on that’s important to you. You talk to ten clients and you have some basic KPIs that are the same, but you have a lot of that are very industry specific and our customer specific, hey, this is the KPI care about and then think about in the organization what’s the CFO’s KPI compared to. Let’s say a call center manages KPIs, they want to Mount it. So it’s very exciting because you’re in different points of view and a lot of times everybody is correct. And I love it because even though I think we do a great job and we are doing a great job, we’re also ourselves learning on a daily basis. And that’s what makes it so exciting to be in this business.
Well, to carry through what you just described. This is interesting because your outcome driven as any company is whether they know it or not. And the outcome is often some measurable KPI, like it could be average call duration reduction by 30%. That’s fantastic, right? But that’s actually a side effect of what you’re doing because your outcome is actually empowering the sales flow to generate business. And it’s this weird sort of intangible thing that has a lot of tangible measurements along the way. But because that is your outcome, you’re building the product to drive the business there versus drive call duration down.
Correct. And again, some of these things happen without us even focusing on, for example, getting that average handling time of a call from five minutes to three minutes means we could handle more calls per day, more calls per hour per agent. But the other great thing because we are AI and we don’t have physical bodies in the seats. For example, my earlier experience with that airline with a two or three hour hold time, nobody wants their customers away. Even an hour on the phone, usually two minutes or three minutes is kind of every business is different. But here’s the thing about this with AI, if you said, hey, Brian, I know we have 200 agents on standby, but can we get to 1000? It takes us about an hour and we can ramp up to have 1000 agents just the way our voice works. Or maybe we needed better yet. Hey, listen, the last two weeks of the year are dead for us. Can we just pause and the answer is yes. Or how about this? Let’s say you had the best, most expensive data of all time and you realize the only and we just actually had this with the client who we lost.
We found out with their very good data that they purchased that there was, and we found this out within a couple of weeks just to back up and jump around a bit. Ai does learn as it goes, but you can’t expect changes by minute or even every half hour. You need a lot of data to make decisions. And there’s of AI sales as education. It’s not a flying unicorn here’s where it can do today. So going back to that point is, let’s say I had those agents and I wanted to really optimize my data with this client. We realized there was a two hour group here, then no calls for 2 hours made more sense than another 2 hours. So it’s basically three two hour shifts with an hour or two in between that optimize their data. So not, let’s say, quote, unquote, burning leads, calling them we know had the best results. And where else are you going to be able to hire people for 2 hours, send them home for 2 hours, come back, no one’s going to work a split shift like that. And more importantly, now the customer is getting a call when they’re able to talk.
And so again, customer experience went up and then the company made money. And I find that if we look at that customer experience first and really care about a client, the money usually works itself out. It really does.
Yeah. This is the other thing people need to know, too, right. This idea of outbound conversational voice, AI is very different than Inbound. Like inbound has more measurable patterns, more specific use cases. There’s probably a lot of tools that they’re trying to measure data that’s doing post call recording analysis, like capturing sentiment analysis, doing certain things like that.
Tying it into customer record. There are certain things there’s stuff that’s very much about enhancing the data and customer knowledge, not the experience, but the knowledge within the trust that they can use that knowledge through the customer record to empower a call center rep or an agent or a sales rep to then have follow up, which will be more meaningful versus you’re literally. I’m going to go find people and I’m using conversational voice AI to do this. So that then when they’re at this, we empower people to take in this knowledge. And then you can do the hand off right back. Like, let me grab somebody who can take care of this question you’ve got and then immediately handle a person that they’re warm, they’re nurtured, they’re engaged. That’s awesome.
Yes. One of the other neat things we do, Eric, because we’re a turnkey solution. We’re not just saying, hey, we’re the AI brain. So anything that’s telling VoIP related, go talk to your vendor. We’ve built this all in one package. So now again, making sure that the customer experience is best regardless of AI. But because we control the technology, I could do things like this. Hey, Eric, I see you’re calling back in. Last time you spoke to Brian. Would you like to speak to him again? I don’t need AI to do that, but I need to handle that at the switch, right? Yeah, but we’ve also had some pretty funny experiences. I had this great message from basically when we first developed this, I said, yeah, it would be great. Let’s go help a whole bunch of hedge funds and VCs and all. And they’re out there trying to raise money to be able to fund companies. Let’s use our technology to help them open accounts to raise money to then invest in us. So one of the first companies we were trying it with, he sent me a message and I love it. I saw that on WhatsApp it’s screenshot of my LinkedIn.
But basically he said, hey, I missed a call. I was on another line, I called the individual back and he said he wanted to speak to the other guy. I explained to him that was AI. He said, no, I want to talk to the AI instead. So the guy got so frustrated, unfortunately not understanding it was AI, hung up on him. And so we’ve had some interesting stories as we’ve gone. It just shows how conversational and humanlike that this experience is. And remember, artificial intelligence, all it is, is a way for computers to think and act like a human. That’s what we’re trying to do. And of course, you mentioned earlier, and I love those if then else statements. I always see that’s all AI is if the L statements. I think I could kind of make that argument to a degree. But it is a little more than that. It’s really understanding interactions with people to make them feel comfortable. So as great as the technology is, the scripting of it and how we implement it is just so important.
Yeah. The last thing you want is like, hey, we’ve got this beautiful framework that you can do and you’ve got to learn now, VML, some new markup language, some new proprietary thing, or God help you, it’s written in JSON or some bizarre coder unfriendly language. So you basically need to become a coder and you need to hire a team of experts. Like, I love the turnkey approach, because that’s really I’ve looked at a few different ways of even like rudimentary, like call flow handling for stuff. I use a product called Go High Level, and it’s very good at SMS based capture and it routes it to your phone. So I can enter into a text conversation pretty quickly and I can build some very simple upfront stuff. Nothing more than that. Right. However, there’s lots of other third party things that like, we can take that call, that initial SMS, and then we could route it and we can add more knowledge to it. But then I’ve got to train this bloody thing, integrate it with my tool. And now I own the relationship between these two techs, and I own the programming of it. So I would never go down the road with it.
It’s just not in my best interest to spend the effort there because the ROI for me is not there. But if it was, I’d be like, let me give it a whirl.
And that’s the fear of a new technology. What’s the cost to learn it? What’s the cost to understand it? What’s the cost to implement it? And then if it doesn’t work, I just wasted time and money. And on top of it, that’s what stops a lot of technology going forward. Or when they hear of AI, that’s voice Basic, first they think of an IVR. Or I’m sure you probably thought of this to the all auto warranty calls that we get, which are actually Robocalls. Right? So there’s a lot of confusion on a new technology, and it’s very common for us when we learn a new technology, you want to compare it to something, you know, just to help make it understand better. So when we were first going down this path, kind of going back to what you were talking about there, these are some of the challenges we had. So I said we’ve got to be able to be turnkey. And I’ve made very success in the past saying, I’m your neck to grab. I’m not going to finger point. So you’re going to give me everything after we’ve done our needs analysis, realize we are, we can’t help you, we’re going to take it from you.
I don’t want to talk to your tech guys outside of getting me recordings. I need get me a couple of phone numbers to transfer calls to, that’s it, and then we handle it all. You’ve probably also seen, like, little things again, outside of AI that we’ve had to overcome. I’m sure you’ve gotten phone calls before. That’s a scam likely on them. Right? So the D ID management and rotation and all of that especially when you’re doing the good things and you’re doing things right. But somebody spoofed your number from an offshore call center or one of these robot Islers now is a problem for you. So when I say turnkey, we handle the caller IDs, the reputation management of those caller IDs, the routing the calls correctly. Based on things like that, we are implementing things such as voice print signatures and all and all of it to make that customer experience again. Better. Again, if I get a phone call yet to the point, tell me what I want and leave me alone, right? Nowadays people attention stands are shorter and we want to be able to encourage that to be as quick and as efficient and accurate as possible.
I always think if you ever seen the scene, it’s from a movie called Boiler Room and Giovanni Ribisi, he gets the phone call from it’s like Wall Street Journal or New York Times. Good morning, sir. Do you like to get and he’s like, no thanks. And then the guy’s like, he goes, okay, thanks for your time. What? That’s it? And he’s like. And he walks it through this whole thing. He’s like.
He gets all fired up. He’s like, yeah.
And he goes So do you want your subscription? He goes, Hell no. That was great that the disengaged outbound call is worse than not getting the call 100%.
Let me tell you, Eric, I can’t tell you how many times just over the last few months that I’ve been interested in a product or service or maybe a JV I’ve seen as an opportunity for a company where I filled out a lead on a website and a week or two goes by. I sometimes don’t even get an email acknowledgement. So sometimes I’m like, I got a little thank you. And by the way, little side note, I always hate when I fill out a form, I just get a thank you. So even on our website you’ll notice if you fill out the form there’s a video pops up from Jimmy and I say thank you for your time, right? Yeah. And acknowledging people when they give you something to say thank you. And so sometimes during our AI process, if you’ve taken existing processes from companies we’ve been able to consult and just try to make that process better. So the AI comes out that much better. So for example, if I called you and I said, by the way, how old are you? Are you 65 or under? Let’s say you say you’re over 65.
A lot of people go, okay. And they keep going down the script, then they acknowledge that they just got this information. How old are you? I’m 65. Okay. Thank you for that. Just little pieces of being polite outside of being conversational AI on the technology. But just sometimes how to be a good listener, how to treat people with respect makes that experience better and that’s outside of the technology, but it complements it very well.
Is there a customer or a type of business that you want to begin with because of whether it’s complexity or sector, business or risk profile because of that engagement, is there anything that you have to watch out for in your line?
Well, we’re very ethical companies. So of course, any business that’s, shall we say not in an ethical or sustainable scalable model is something we’re not going to work with. We’ve actually had some very interesting requests from international. So, for example, in Europe, where online sports betting and gaming, there is no regulations here in the US, if I provide my technology to Europe, what happens? Right. So there are some legal things that we need to kind of ferret out and see how we’re going to go where our typical engagement is. I always like to believe in a quick win. Right. There’s nothing worse. If you’ve ever done a long technology upgrade and you start today and six months later and you’re waiting and you’re waiting, people lose. They’re like, forget about it. They’re not as excited about it. The support behind it’s not like the beginning. So I always like to get what I call a quick win. So our typical engagement will say, hey, listen, if you’re doing an outbound sales process, let’s not do the entire sales process. What are the three to five questions we need to do to qualify somebody and now transfer it to a live agent so that they can be actually more time pitching and selling instead of prospecting?
So let’s do that first. Now, phase two, let’s ask two or three more questions and maybe steps four or five and six. Now let’s integrate to your CRM. Let’s integrate to your payment processor. But let’s try to get a quick win so that within 30 days of engaging us, you have something up and running and you’re starting to see a measurable result for the nerdy Geekies out there. Think about agile. Right? That whole framework, the manifesto. How many times have people like people have been working on my system? Six months. I see nothing different. All of a sudden you had a red button that took ten minutes and wow, look how great the system is now, right? So sometimes it’s not what’s best technology wise. It’s getting a customer. That’s where the business and sales side comes before the technology. That’s why I like to say we’re a sales company. First is let’s get a quick win within 30 days of being engaged. So you can turn this thing on. See that it’s making a difference to your organization, to your customers. And then over the next six months or a year, this evolves kind of like any business.
Right? But the nice thing is once we get that first process in and set up now, we could start talking about others. And it’s also outside of getting those quick wins, a lot of times during that implementation because it was just a small piece. People understand the technology better. So, for example, if you’ve ever bought a car that you spec out online or you ever had a house built from scratch, almost invariably somebody say, well, if I had to do it over again, I would have changed this or done that differently. So trying to plan a full AI implementation of business can be very time consuming. At the end, people go, I wish I knew this. So by doing a small, quick implementation, getting something up and running, and now understanding gives a better idea of how we could build this going forward so we don’t have to go back and redo things.
How do you handle dialect accenting? There’s a lot of linguistic challenges, especially when you have global audiences. You’re serving multiple GEOs. I mean, some people even make fun of me because I speak Canadian, not English. Right. I still say process and project instead of the proper way. As I’m continually reminded my wife is American. Our poor kids are going to be just so broken. They’re going to sound half from New Jersey and half from Toronto. And it’s going to be a really strange mix.
Sure. Yeah, it’s a great point. So right now, we’re really focused on English. However, if you do have that, say, a thick Creole accent or if you ever met somebody from Louisiana and you love the accent, if they get going quickly, I’m like, what did they say? Right? My wife is Costa Rica, and the first time we went to the Bahamas, which we had a lovely time and people are lovely. We went to the little shops there, get a little knickknacks for the family. My wife’s like, what language is that? I’m like, that’s English. Right. So even though we heard it and understand it, she had a hard time. So the bottom line is the AI, which is basically using speech to text if you use your phone and you’re talking to it or using voice based type searches, as accurate as that is, we’re very accurate on it. Now that’s the only piece of our technology that’s not proprietary, by the way, is the St T. We use a product by IBM’s Watson, by the way, and great product, but it took us a long time to figure that out. But it does learn and get better over time.
And it’s really interesting. It’s more than the absence of the issue. Eric is like when we get to a vertical, for example, medical, we have all these different products or medical terminology. That’s where a lot of times training those models ahead of time is just so important. Again, having that 500 to 1000 recordings that we can help train the models makes those implementations much more efficient. But the accidents we do pretty good with, we are trying to work on where the outbound AIS more match the user’s accent, but it’s kind of hard on an outbound sales call because you got to have something kind of preselected, so to speak. But over time, we are getting very cool. Yeah.
Those things of the nuances of an industry and even just we have the curse of knowledge as humans.
So we throw acronyms around. And the difference between an IVR and an IV drip are fundamental. Right. But IV could have multiple meanings, right. There’s like lots of if you go from one place to another, even the same acronym can be meaningfully different between industries and you have to be able to adapt to that. So that’s a unique challenge of doing this stuff. So you folks are doing pretty cool things.
Well, one of the things what we’ve done is and Jimmy, who’s my partner. And again, there’s no vocodo with that. Jimmy. It’s amazing the brain he has and how he came up with this. So we started going down and you’re talking about the acronyms. And by the way, why is acronym such a long word? But I digress. But sometimes you hear it on a call, you write it down and look it up and you go to Google to search it and you’ll see 20 different meanings for it. So for example, we call actually our AI Adisa, which is a digital intelligence sales agent. Well, Diesel has already been taken by the military and a whole bunch of other people. So Dice itself has a lot of different meanings Besides what we use it for. So it’s really hard to come up with an acronym that probably nobody has used before, which is very interesting.
The best thing you could have is when you Google it, as long as Urban Dictionary isn’t the first response, you’re in. Good.
You know what? There’s probably a different type of podcast. I would go into that further with you, but yeah, so we’ve had those before. The other nice thing what we do is we try to even though we have our main conversational model. So think about it this way. Conceptually, in any call center outbound campaign in the world, you’re going to have common rebuttals when I call you. Who is this? Why are you calling me? How did you get my information? Let’s say the top ten. I’m not interested right off the bat. And you have these top ten. But again, as you talk about industry or vertical specific, like if we’re in legal or we’re in home services or whatever, you’re going to have different acronyms and different learning. So even though we have our base model for, let’s say, general conversation, we also have our models per vertical that get better and better. So as a company, we decided we can’t be in every vertical at the same time because we have too many models. We’re trying to train. Right. So we try to be within three to five vertical tops at a time. And until those are up and running 100%.
We don’t want to go outbound to other verticals just because we know what time it takes for that model to be trained. But the great thing is once we have, let’s say, a health insurance type client set up and running, we now have a model train that allows us to go after other insurance type products, whether it’s wellness checkups from a doctor’s office, appointment verification for a hospital, health insurance quotes, benefits verification, which is a huge call. You know how many doctors call a health insurance company just to verify benefits? Last time and again, you’re in Canada, you don’t have this issue. But as Americans, we’ve all been a doctor. In fact, I just had an appointment where I already had a Fax in there. They said Fax, which was great. I took a picture sending a copy of my insurance card so they could call in to verify benefits. How is it in 2022 that people have to call in to verify benefits? Because you’re calling in to somebody’s looking at a screen. But however, that’s another problem we can help with.
Yes. Even when you get, like, interagency stuff. That always drives me a little bit nutty when you think of the government work. I used to always get it when I would fly. And you have to be careful how cheeky you get with those. The folks at TSA, because they have to be a little bit more bland. But they would say, like, when was the last time that you crossed the border? And I’d be like, being honest, I fly a lot and I can’t remember. And you’re probably looking at the date right now. So whatever that is, it’s right about three weeks ago.
Yeah, for sure.
It’s like that data is out there. We shouldn’t have to have this weird hand offs and being able to introduce it, capture this data, and then feed it to a system that then can be built knowledge for a human to take it up from there. That’s the next Goldilock stone as far as I’m concerned about where conversational AI needs to go. It’s not just in treating it as a one and done, but taking that knowledge and then taking it further into enhancing the overall customer experience. And ultimately, even though we’re talking about enhancing the sales flow and business flow, you do that by enhancing the customer experience. That ultimately is always the true initial goal. The end goal is that what does that generate? More sales, more happy customers, more post sales, customer success opportunities, stuff like that?
Yeah. I always like to say our technology is here to help you sleep well at night, Mr. Company Owner. So we were like, hey, if I get 10,000 more sales this month, I need to hire another 2000 customer service agents or whatever the numbers are. Well, now guess what? With AI, we could just go ahead and ramp it up. Or if you’re a seasonal business say I sell flowers. And during Mother’s Day, wow, my business goes crazy. So maybe I need 5000 diesels, but after Mother’s Day, I only need 50. Well, again, I think Kobe taught us all. You got to be flexible in business. And so we allow that in our agreements. Like, here’s your minimum, which could be like ten. You go up to 5000, down to 500, and go up and down as you need on even a daily basis. Let’s just say you’re having one of those days. That’s just so slow. But going back to the data, Eric, which I think is an important point. A lot of times when we’ve talked with AI, people are talking about, okay, well, if you know you’re calling Louisiana, you could have a different accent and they think about things like this.
But it surprises me how many times external factors aren’t considered in making a successful contact to a client. That’s what we’re trying to do, right? We’re trying to get somebody on the phone’s, requested information or answer their call if they have a customer service issue. So we’ve tied in external factors. For example, what’s the weather? If Oklahoma just had a tornado, should we be calling people in Oklahoma? Right. However, it’s a Super Bowl one. The Super Bowl here. This is football for you, Mr. Eric, not hockey. Sorry about that. Yeah, well, the week where the team won. It’s been psychologically proven. When a sports team wins, people in that area that are sports fanatics spend more money. If their team loses, they spend less. Well, where should I not be calling? Possibly calling this week as an example.
If you want to be avoiding Cincinnati for a sales cost.
But there’s so much data out there, sometimes you don’t even know what to look at. And we’ve also assisted clients, non AI related, again, because we control all the technology. So, for example, we had a home security company, and their service really is based on your credit score. Well, there’s no way through public information can I get an accurate credit score. You just can’t do it, right, because of Privacy, et cetera, which is a good thing. However, there’s a lot of factors I can get from you called in to give me an idea of your credit. So, for example, if you call me on your at and T cell phone, I said, well, if you have an at and T plan, you probably had to have halfway decent credit to get that phone compared to a prepaid service where maybe the credit is not as high and somebody would need to send them to a different Department homeowner compared to renter, things of that sort. So there’s like 800 data points we have thus far. And I think we get more to, again, improve that customer service to help them get there and also help business owners make better decisions on who their customers really are.
By the way, I was also much more wrong than I am right. So I was just like looking at the data and we have so much data for the first time, you couldn’t get another great example is if you asked me what colors the sky. I said blue. Well, does that help the client better than saying dark blue, light blue depends on the weather. So we can have all these different responses. Think A. B. Testing on the web, but through our AI system to overtime to figure out which are the better results for each client. And that’s where it just gets so much fun and exciting. And one last point, while we were developing this and we were going through the different verticals, et cetera, we realized that the reason we had to develop our NLP was that the NLP that are out there all have a similar way that they process a conversation or the text. And you’ve probably seen it yourself, even on web based chat. You’re texting and you hit Enter and you see it takes like a minute or tens of seconds to come back. And we didn’t have that luxuries.
We had to figure out fast. So Jimmy and I realize is that the way that NLP process information, in our opinion, was an incomplete process. So we’ve come up with our own process and our own NLP of how we process a conversation. And to my knowledge, it’s not done this way. We’ve looked at it, we’ve done patent searches, et cetera. We’re still trying to find the right patent attorney to help us get that one done because it’s such a unique process. But this system has made it so much faster and so much more efficient on how it processes the data that the conversations are much more rapid and much more accurate. There’s examples. For example, if I said, where are you located? Or I said, where did you get my information? There’s a lot of products out there here to see the word where and say, oh, Boca Raton as an example. No, that wasn’t the intent of that sentence. Right. As a human, we got it right away. Other things like if I asked, are you over 65? And you say, I will be next week or I just have my birthday, well, that wasn’t a yes or no, right.
What did you mean by saying that in this NLP that we’ve created to be much more conversational, we took into consideration a lot more. And that’s why I think it took us an extra year to get to where we are today. But I’m glad we did because we have such an amazing product and service today. And more importantly, even though it’s great, we’re meeting so many other companies that are trying to implement AI that have other ideas and thoughts that they have that helped make us our product or service better. But at the same time, those conversations we have with other owners has just been exciting, too. And every 90 days, right. There’s so many neat things happening in the industry that it’s exciting. I mean, we really are still at the very beginning of AI. We’re definitely not halfway through. Regardless what you see online, the term AI is thrown around everywhere. Everybody’s AI now. Sometimes they just add extra fields. And I’m like adding extra fields to a CRM. Sorry, Mr. Sf out of San Francisco. There. That does not add AI to me. Help me make my business better, help me be more productive quicker.
Don’t make me analyze the data. Just give me the answers. And that’s kind of what our approach was. Yeah.
And that thing around latency in the discussion, I even had to just even funny. A visual audio matching is a real challenge that I bumped into early on. I actually changed platforms and changed some other hardware that I used for the podcast because I was using a really good streaming box that would go in and then I would have my USB microphones. I had a really good microphone and all stuff. And it was like a 200 millisecond latency from the audio channel, which it wasn’t obviously noticeable, but it was enough that people it kind of makes you squint a bit when you’re looking at someone talking and their voice and their audio and their mouth are not quite on, but they’re not off.
So it takes you out of the moment. And when we get the same thing with the conversation, when you hear a conversation with somebody and it sounds like they’re, you know, the CNN person calling a guy in Cobble going, yes, we’re on the ground here. And it’s really good, really fast paced conversation. I’ve got a great answer. Absolutely. There’s a lot of stuff going on.
Jeff, thank you very much. When you hear those gaps, it makes it mechanized and it removes your ability to stay. So you find yourself kind of like, come on, leaning into and it changes your mood about the conversation at that point, 100%.
Now you’re anxious. What that’s going on? It’s unknown because it’s not natural, right? It’s not conversational. It’s not humanlike. Again, AI one more time is how a computer thinks and acts like a human. That’s the basics, what we’re talking about. So a human is not going to take something and go, I’m sorry, can you repeat that? I’m sorry, can you repeat that? Which gets annoying. What would happen if you asked me a question? I didn’t know the answer. I’m going to say, I don’t know, let me get on Google and I said, let me call Bob and find out. And which Rob was saying that.
Hey, you know what?
I don’t know the answer. But Eric, I don’t want to get you wrong information. Give me a minute and most people go, okay, so if our AI gets stuck and has to transfer it, that’s good, because Eric got the information he needed as quickly as possible and didn’t have to worry about, oh, well, I don’t care about this person having the wrong answer. I’ve had that case where I call somebody and I ask them a question, and I could tell they don’t know. They’re like, you know what? I’m not really sure. I think we had somebody a couple of months ago that has I’m trying to remember. I’m like, I just want the answer right. How many times you have to talk and they’re looking something up on the screen, and all of sudden a you go, what colors of the sky? And what are they doing? They’re looking it up and you don’t even know. Have you ever been on a call you’re like, are you still there?
Yes. Am I still on the call? Right. Because again, it’s uncomfortable now, and now you’re in an uncomfortable, anxious situation. How much did your value drop of that company you’re working with? What’s the chance of you doing business with them? Because that’s an extra piece of uncertainty and doubt. So we want to remove all of that and make it as smooth as possible.
The challenge we often find, too, is when if you’re experienced at calling a call center, you’re actually not the target customer that you need to have. So you and I, we joke about this stuff because we’ve been inside it. I’ve helped build IVR systems. I still remember Meridian Mail mailbox number. I know the voice. It’s burned into my mind. I know her voice. If I ever met that person in real life, I’d be like, you’re the Meridian mailbox mailbox number lady, and we have to build these systems and think like systems. But the people that are really affected by this are the people that are just calling to talk to a human to get a question answered, or they’re getting a call now, an outbound call, and they want to be engaged, and they want to be able to have a path to somebody not to have an outbound call. Do you imagine this, an outbound call? And then someone responds back with, I’m sorry, I didn’t get that. I’m sorry, I didn’t get that. I’m sorry you would end up having somebody who’s like, look, you called me, and then they’d be screaming like Samuel L.
Jackson English mother effort. Do you speak like he’s yelling into the phone? Terrible experience, right?
Absolutely. Yeah. Your stuff you’re doing is Wild Pat Fleet. She was the one who did most of the voices for at and T back then that we all knew when you call, the number you have called is and that kind of thing. And she worked with Susan Bennett, and I forgot the third. I know a lot of telecom nerdiness.
I almost want you to get you to say is because you have such a fantastic voice. I almost wanted to say, like, high ball, three, high and wide, just a bit outside. You’ve got a perfect announcer voice.
You know what? I had a very short lived radio show in the 90s, and we were taken off immediately because nobody wanted to hear what we talked about because it was just me and my friend. And we were just saying, but I appreciate the compliment. I’ve heard it a lot. And I do get told I’m animated, which is when I do the voices for the AI. So in our very beginning, when we had our first recordings and we were kind of showed off and Jim and I said, listen, we’ve gone the organic route in business before. This is not our technology is too strong. Let’s get some money behind us and really push this. Like any business you want to be big, you need some resources behind you. Right. So as we were going out there trying to find investors and we would show recordings, we had people not calling us back. I’m like, I thought that was pretty good. And what we found out was a lot of people thought it was fraud. Well, that’s not an AI. You’re just talking to somebody and you recorded it. Right. So I actually made our first AI is me talking to me to remove that doubt.
But I’m very animated when I talk and can be. And so a lot of people just laughed at it because the AI took on my personality. Like, okay, and things of that sport, it was a little extreme, but it was still kind of fun. But I appreciate the compliment. Thank you.
When you go that route and that’s interesting thing. Right. Like the bootstrapping and revenue funded and that decision to go the next phase of growth, especially because you’ve got a strong amount of technology investment you’ve got to make. What was the sort of driver that said, we’re ready and we need this?
Yeah, there’s a couple. Number one, it took us about a good year to get past the latency issues. It was just too slow. And of course, we were trying to do things like where you cache the recordings or cache the voices and you go through all these tests. And actually our first one that we tried to put into production with a friendly client where we just wanted to test, Jimmy said, okay, listen, this works great, but it’s got to catch the calls. So your opening statement to an individual has to be at least 7 seconds long so we can load all the data. I’ll never forget it’s like, hey, my name is Brian calling with XYZ. So how are you doing today? We actually put the Oz and whatever. Actually, another guy worked with us, did the recording, did a great job, and just vamp. And so the funniest thing is that initial thing was saying, how are you doing today? That’s a human thing. You’ll never hear an AI really do that. And so in our AI, because we have things like how you doing today, to be more personable, to be more human was why we had an issue in the beginning, because they said, well, why is the AI doing that?
But once we had that first client who started seeing results, started getting the calls, and that’s when we said, okay, to go from A to B, it’s going to cost a lot more than we thought. And unfortunately, AI processing all is still not cheap. Any engineers or developers you need that have that skill set. And remember, when you’re in a brand new industry and AI really is brand new, you’re not going to find people with ten years of experience in AI. You’re not going to find people with five years experience. So you got to really kind of break it down to the piece, hey, we need a react guy here. We need some.net people there Cosmos people, whatever it is. And so we had to break it down. Well, all those people cost a lot. So in the beginning, Jimmy and I’ve been very blessed that we had enough friends and family to actually work with us knowing that, hey, we’re a startup, this may crash your burden, you may never get paid, but if so, here’s some stock. But we got past that where people are like, well, this is great, Brian, but I still have electricity to pay, right?
I still need some money coming in. So that’s when we said, let’s go ahead and raise that first million dollars. Let’s get out of the garage into an office. Let’s raise another 5 million. Let’s get the product done. Let’s get into revenue by this first quarter, and we are going into revenue this first quarter, which is great. But now our next thought as a company is it’s a lot quicker to buy revenue than generated. Remember, not everybody knows conversational AI for call centers exist, so it’s not a lot of search results for it. So there’s a lot of educational sales which traditionally cost more. So what we’ve done is we’ve been identifying targets and trying to buy complementary technologies to make our product better and get to pieces quicker and also add revenue because they also have a customer base that we can sell into. So it’s been an exciting time. But our next phrase is definitely more about acquisitions than implementation just to get that revenue quicker to get our evaluation higher. But the initial one was just like, we need a few good people and they’re not going to work for shares anymore.
They have some bills they need to pay. And I’m really glad we’ve done it. We’ve got such an amazing team. We’ve been again, so blessed, I can’t explain it luck, whatever you want to call it, that we’ve been able to find the people we have to work with us and take a risk as a start up. And I think it’s starting to pay off for a lot of people though. It is a fun business, too. I think that’s what makes it interesting is a lot of developers take a job and I built a database, and there’s really nothing you could show friends and family. Right. With our product. Every time somebody does something like, hey, look, there’s Brian doing another call or, hey, I’m Brian speaking over here, and I’ll do a live call and things of that sort. By the way, that’s the scariest thing. When you do live calls in front of like 1000 people or 200 people and those technology Gremlins love to come out of the woodwork. We’ve been very good on that, too. Yeah. So that’s always the scariest moments for me. But the technology is so strong and the people that we have working with us because again, we were able to raise the money to get some very good engineers on here, and now they have a product that they could see and touch and show off.
I think that’s what gives them the excitement about working with us, too.
Yeah. This is the real leap of faith is two sided in the startup ecosystem that people don’t realize is that number one, as an engineer coming on board to a company that specializes free revenue, there’s a real challenge of like, is this the right place? I need to be. Everybody’s got kind of there’s a lot of good ideas out there. There’s a lot of ideas. Then there’s a subset of those that are good ideas, then there’s a subset of those that are good ideas that have revenue opportunity and potential close at hand. So the field, like any marketing funnel, narrows. But then as a founder, you also have a leap of faith of you’re making an idea that you can bring to market that, you know, you have a path to revenue. You know, you have a path to growth that when you bring somebody on board, it’s a beautiful relationship. Those like early employees where you learn so much from each other, and it truly is like a family, which I know gets overused. And in fact, people get sometimes really angry when you say, like, if your company says that it’s like a family, it’s because they’re going to abuse your hours of your week.
Like, that could happen. Maybe. But I found quite the opposite. Like, I started with a start up and I was like, employee 200. So it was a big family already.
But we grew and grew, and I saw that culture maintained with growth, and it’s challenging to maintain that culture, but that two sided faith continued so much through it, which was a beautiful thing.
Well, we took some chances in the beginning. Again, I used to work team more than family because I always want that in the concept. But 100%. But some of the things that we did in the beginning, which was looking back now, I guess was a little gutsy in hindsight, is always 20,020. But one of the things I’m glad we did from day one, as soon as we started hiring people and bringing them on full time, saying your vacation time is unlimited, if you feel you need to take off two weeks or four weeks or whatever time you need to take off, I’m not going to give you two weeks of vacation pay. Take as much time as you need. Keep in mind, we’re a startup. There’s deliverables as long as you need them, do your job. So we took a little bit of chance to give them people to make sure that they understand that, yeah, this is a company, but I know your family at home is important also. So if you need to take off that week or two, I want to encourage it. And if the company has got to redo things around that, we’ll do it.
And a couple of investors in the beginning are like, are you sure about this? And we’re like, I know it’s a start up, right? It’s a scary, scary thing to do. And then also because again, we want to encourage to get good people, our cheapest employees. The bottom line, we have nobody that’s under $35,000 a year. And we could have gotten some people that were cheaper than that. And I saved a little money. But again, even like our base receptionist up, we want to make sure everybody that came to this company realize that we do care about your time. We do value your time. And the other hard thing we did again as a start up, a little gutsy was we got everybody. And again, you’re in Canada, you don’t have this issue, but we got everybody insurance from day one. So we raised a little money. We had to find those investors who agree with those values and not just the technology, because as we all know, it’s that team, those people that make something happen. And if you’re having an issue with your AI or you want to develop something, we want to make sure that that person who takes your call as our customer has a great experience, too.
And how do you do that? You got to take care of your people. So we’re very happy to say that every single employee that works for us and I don’t care what they do has shares in our company. So if we were on track for a Nasdaq IPO June next year, and if we get bought beforehand, great, but whatever. But worst case scenario, by 18 months or so from now, everyone’s going to have a nice little payday ahead of them, right? So we wanted everybody to not only have a good working environment and proper money and benefits today, but also have to have a future just like Jimmy and I are.
When it comes to road mapping and execution, I’m curious, especially because the complexity of what you’re building when you look back across road maps that were 18 month road maps. How well have you been able to stay on top? What are some of the biggest sort of things that you slipped, but for the right reason.
We basically had to redo our VoIP switch. So before we had Omar working with us, we had another gentleman who did a great job. But again, remember, we were startup. We didn’t have cash. So this was like, Jim and I have our own pockets paying the subcontractor here or hey, I owe you one kind of things. And our initial switch did get us a proof of concept. It did get us our initial funding. And then Jimmy knows me pretty well. And the last couple of companies we worked with, I’ve outsoled production. I know you’re shocked direct to hear that I’m a sales guy, but I’m a salesman. So he said, hey, listen, with this switch, we’re not going to be able to grow that quick. We’re not going to get that many businesses. We’re going to have to have a whole bunch of clusters. And this is not going to work. I think we need to start over because again, it’s such a new technology, what do you need to build and how to build it? You don’t know. Going back to my early analogy, you have a house built, you think you know everything. And then once you’re in, you’re like, man, now that I built it, I wish I could have done this differently or that differently.
Well, this was brand new technology. There was no resource to say you should build it XYZ. Nobody said the foundation should be six inches deep or anything of that sort. So after building it and realizing the latency that was happening on the void side and getting the privilege of working with Olmer to join our team, we had to redo that. That threw our time frame off by three months. There’s a rule of thumb. They say when you do, a technology company usually takes three times as long. It costs three times as much. The good thing is we came about a million dollars under budget. It took us about an extra five months to get into commercialization. But now we have something I could scale before if you would have said, hey, Brian, I have 50 days. I need to go to 500. You have to give me three days notice. Now it’s literally an hour or two again. So I’m glad we did that. We took the steps back to take the steps forward. And now that we have the diesel running, actually this week we released a 2.0 product, which is just so amazingly faster.
And we already have the 3.0. But that was our biggest, biggest change that we had to do was that VoIP piece. And it was just so underestimated by us. I guess in hindsight, being 2020, we were so concentrated on the brain and to get that work. And when we got it done. We’re like, yeah, then I’m like, cool, you plug it into Twilio and it’s like, no, I’m like plug it in and let’s make some money. So we had to take that Dag. It definitely cost us a little more. But again, we still came in under budget, took us a little more, and now the product is just so enterprise ready that we can’t scale up. We also redid another piece of the architecture, how we do the administration of the product. So before we had, let’s say in a typical CRM or most software, you have like say a God level and then kind of filters down to your managers and agents, we had a similar architecture. But with Jimmy realizes, hey, what if there’s a hiccup in the admin level? It could affect each client. So not to talk nerdy geeky, but just conceptually each client in their own basic silo.
Even our administration piece of it is a separate silo. So if any of these silos were to crash or have a problem, the rest of the machine keeps going. So we can literally have the administrator disappear and each client would be up and running. So that took some re architecture on the Azure cloud. And Jimmy and those guys, they could talk much more detail than I can. But the fact and we’ve actually tested it. We’ve done over a billion phone calls of the simulator and we’ve stressed test the heck out of it. And the thing is just cruising. The other thing about redoing it and moving from one cloud to another, we wound up on the Azure cloud testing. Our cost came down dramatically. So now our margins are where you would expect to be in a SAS model. And more importantly, we could scale so much quicker, we don’t have to order a lot more processors and memory so we could scale without any forklift upgrades, to use an old IBM term.
There you go. The one thing is there’s a lot of technology required in order to make technology be more human, right? And the ability for AI like now at least the accessibility of these technologies being on demand, seeing that commoditization of the back ends huge boon for you because those providers are never going to do what you’re doing. They just can’t. It’s too close to the true customer value. They’re in the infrastructure game, they’re in the services game. So this is a big win because ultimately Azure is going to love you, your customers are going to love you. You’re in a two sided marketplace on both sides. And it gives you as you grow, the opportunity to get wins with the provider side. And then if that the economy of scale starts to happen. And ultimately, then the customer, the true winner of all this is that end customer end. Like all the way. The person that’s getting that outbound call from that company that’s using you, that they don’t have to know you’ve got as your digital twin that’s making something happen. And so many layers down the line, they don’t need to know that.
But every layer has one because of the advancements in technology that in the end made the human experience better 100%.
And the other exciting thing on our roadmap is even though we focus on voice today, I’m sure you imagine, oh, well, I could also get a text, right? I could also do this through a chat or messenger, WhatsApp, etc. Or. Well, those conversations that you have over text and the conversation you have over voice are different. An email you get is usually not how you would speak to somebody when you would get an email. So how we converse going over different mediums is also a challenge because if you got a text that sounded exactly how I was speaking to you, it would look weird and vice versa. So the other exciting thing about our technology as we’re going on the channel is how to have the conversation adapt based on that medium you’re on. So again, to make that customer experience better. And by the way, Eric, I may call you and you’re like, I don’t want to talk, but can you give me a quote or text? Why not? How come I can’t? And maybe there’s some sensitive information you wouldn’t want to do, but I can’t tell you how many people are like, hey, let me just text you my credit card information.
I would never do that. But an end user might. So the technology again, still in its infancy, still such a long way to go. But where it is today and how it could be impactful today just to us is such an exciting thing. It really, really is.
It is an exciting time to be in this world and being able to do what we do. I feel blessed because of the opportunities we’ve got and excited by the ability to create those opportunities for others. And you guys are doing great stuff. Like I said, your energy is infectious in the greatest way because it will come through in product. It will come through in telling your story. And I really want to thank you for taking the time to share your story here for folks that do want to find more, which they should, because this is a really wicked cool, as my Boston friends would say, wicked cool product. So Vicodia will have links down below. And Brian, if they want to get in touch with you, what’s the best way to do that?
Brian at Vocodia and that’s Brian with an I@vocodia.com.
That’s perfect. One thing I remember is like every once in a while you’d get these little things to be like a little Easter egg that would come up. I often think about doing this with like a voice AI, because you get those ones. It was like the old IVRS and some would say like, phone this number. It would show up on like Reddit or somewhere and it’s like for English press one for this whatever. To hear the sound of a duck, press nine.
I remember that I used to love the Aflac one we used to call in just to hear the duck or one of the vodka companies had Smirnoff. Remember they had Jacob smirk. You can call in and get a joke every day. We actually have on the new website coming out. There’s a couple of cool little fun things on there that we will talk about. We like for people to find. Yes, absolutely.
That is the fun part about mixing technology and human experience and you folks, you’re doing a great job so Congratulations on so far success and I wish you can’t continued success as you grow.
Thank you so much. I appreciate you having me.