Artificial Intelligence

Welcome to Building Better Developers, the Developer Nord podcast, where we work on getting better step by step, professionally and personally. Let's get started. Well, hello and welcome back. We are into a new interview. We're going to be speaking with Chris Barkhurst of Barky AI. And as the name may imply, we're going to talk about artificial intelligence. We're going to talk about the elephant in the room, as it were, chat GPT. We're going to talk about how to really take this stuff and do something with it. We've had a couple episodes where we've talked about AI, but this time we're going to get a little bit deeper and actually get a little bit of history about it. Because this is a guy that has actually been working with it for a while. Now, I don't want to steal all of his story and take away from some of the background that he's going to give us, particularly because he's lived it, he knows it better. But be ready, maybe take a few notes and think about how you may be able to apply AI to your job and your tasks. But let's get into our interview with Chris. OK, today we're starting a new conversation and we're speaking with Chris Barkhurst. And by him, his description, he stands as a paragon of innovation and leadership. So I don't want to undersell that at all. That's a great word to use. The big thing is, he has spent a lot of time, has started young, as we'll find out, spent a lot of time in data and AI, analytics, all these things that we sort of have our hands in on a daily basis. We've got all of this data and all this information. So I don't want to take away from him as somebody that has found ways to make the most out of that and turn that into a profession, a career, some businesses and things like that. But I won't take his steel, his thunder. Welcome to the show, Chris. And why don't you go ahead and tell us a little bit about yourself? Thank you. I appreciate it, Rob. Yeah. So I'm Chris Barkhurst. I'm the CEO of BBC. We are a parent company of two companies right now. The first of which is our longest running company is Barki AI, which is a virtual call center agent. So when you call up, you speak to Barki AI, an agent, a virtual AI powered agent, which runs on our proprietary AI platform. And it's a seamless integration, very conversational in nature. Excellent. And then our second company is Benson Lloyd Card Grading. And what we do there is we use AI to basically grade cards. The sporting card business is a billion dollar plus business. And we use AI to help spot counterfeit as well as use it to grade cards while removing bias. And then we do a lot of cool stuff with that as well. So it's been very, very fun past five years. Both of those are really interesting and very different, I guess, on the surface, although obviously you've got that AI at the core. I guess to start with the Barki AI, how did you come into that? How did you decide to go into that focus as far as AI from a call center point of view? Certainly. Well, let's take a step back. In about 2022 or so, actually longer than that, 2020 is last year. Excuse me. In 2015, I'll give you my dates straight there. In 2015, I started doing more consulting. And what we did was we would help companies optimize their strategies for long data, stuff of that matter, how to do better analytics, do better visualizations, project management, etc. And right then and there is when the GPT-2 model was released by OpenAI. And we started tinkering with that, creating an application which allowed you to take the company's data into the GPT-2 model and then integrating with companies information like the project name, who the project manager is, what the project is about, the information about who is responsible for what, what code platform should I use, what's the status of the code, etc. So, when somebody was on boarded, they could simply access the system, this platform and start typing away. It was just chatbot. And onboarding was so simple because you just say, who is responsible for this? And the chatbot would respond, John Doe was responsible. And how do I start the code for X, Y, and Z? And it just said, here's a code base for X, Y, and Z. And you're like, this is great. So, we started that with GPT-2. And then as we got more and more clients, more and more clients wanted that application more than they wanted our consultation services. So, we started migrating into more of that application development. And then all of a sudden GPT-3.5 came out, which is basically chat GPT. So, once the world got a hold of that and started developing with that, they're able to start writing stories about, you know, in the style of a pirate and all this other fun stuff. Our clients were like, hey, we've already had access to this for a long period of time. This is nothing new to us. It's just all of a sudden, millions of people are able to start writing all these creative prompts. We just weren't aware that all of the creativity was behind it. So, we started looking at it like, OK, well, all these companies are about to fill a void. How do we get above the curve? And I said, you know, the only way we can get ahead of the curve is what's a void? That's something that people hate doing. Our thing was, well, people hate waiting on hold. And I always hated waiting on hold to get my password reset. So, our developers found a way to do text to speech and speech to text very quickly. So, we were able to harness the power of the Barqia AI system to get on telephony and listen to a phone call, make a quick decision within the Barqia AI system, and respond. So, that's how we did the Barqia AI virtual agent. Does it actually have, can actually trigger actions and things like that on the back end as well? Oh, definitely. Definitely. So, if you're calling up, for example, we have somebody doing customer service for retail location, and you're trying to figure out, hey, it's item ABC in stock. While you're saying that, our Barqia AI system is proactively checking if that item's in stock. And within just milliseconds of you stopping your question, we are already checking if that item's in stock, formulating a response, and then we will speak that back to you, saying if that item's in stock or not. So, we're performing actions in the back end. So, if you want to change your address, or if you need something refunded or whatever, we're checking the back end to check those policies to see if we are able to do that action or not. And if we can refund your account, we can. If we can't, we can tell why not. And all those things are done in real time. So, by the time you're done with that phone call, you can be assured that whatever you called in about has been resolved. And, you know, we have a one call resolution of 1.2. So, you only call in, all of our customers have called in, they only call in one time for every 30 days. So, I think that the average is like 1.9 or 1.2 something. It's something relatively high. People always call in twice every 30 days to get one that you resolved. It's very high industry standard. So, we're very low. Because once you call in, it gets resolved the first time, we never have to call back to the customer. How long has that been? You've been doing that since, I guess you sort of start, so just the last couple of years that that's been out in the wild as it were and being used? Yeah. Well, Barkey AI officially launched back in April. That's when we actually went to a public production environment when we officially launched. We were in a, it was probably called like an alpha or beta state with just a few of our call center partners, just making sure that nothing's getting loose, nothing is, there's no loose bits anywhere. And we took a little over a million phone calls just to test the waters. And of those, just a handful of people understood that, hey, this is a bot. And it was less than 100. But everybody else had no idea they were talking to an AI. They thought this was an actual person because we have infinite voices. So you call up a million times, you'll get a million different people. You never know that it's a different, it's the same AI, but it's a different person. But they all have the same access to your tools and your background and your history. So you may say, well, yesterday I talked to Joe, today I'm talking to Katie, but the AI can see your entire call history, see your entire call transcription and see what you said and what you did not say. So we can actually do all that. It's amazing technology. That's really cool. So is it a, what's the, what's sort of like the level of effort to go to roll that out to new call centers? Is there a whole lot, is it more of a configurable? Is there a lot of coding to make sure that you integrate with all of the pieces that you need to integrate with or what kind of, how, how short or quick or short or long is the, is the rollout for something like that? Well, we can, it depends on the level of effort that the call center or the organization wants to take it. I mean, if the organization wants us to just put it with their policies and just run against the policies that have a very entry level system where we just, just do like the tier one things, like making and changing, adding and removing, changing addresses, doing corporate policies, like here's your, here's our hours of operation, et cetera. And then if there's anything that needs more complication than that, we pass it off as what we call a tier two, then that's, that's pretty low effort, low effort, consider to us. But if you want something that's more complex where we're doing pretty much the entirety of a, of a, of full FTE, then that is more complicated as when we get into the nitty gritty and we actually go into your systems and we help, we work with your developers to figure out how we can wire, you know, when somebody says this, we have to fire this, this action and that's action has to go into your system and pull this lever and this lever and this lever. So we have to make sure we do this. But we have to also, you know, make sure that what they say this instead, it doesn't actually pull this lever and also, so it's very, it's very micro management. We don't want to pull the lever when they say one thing or another. So we will, we have to, you know, be very cognizant about that. Yeah, that makes sense. And that actually was my answer to my next question. I was wondering whether you do like a, you know, tier one, tier two, tier three. It sounds like you can, you can do it all depending on what, you know, how much they want you to, how far they want you to go. And then of course, you know, it gets more complex as you get into higher tiers of support. You know, there's a difference between like swapping an email address versus, I don't know, spinning up an entire server and, and checking IP addresses or something like that, depending on what, I don't know, what, whatever your business could be or, you know, tracking orders and all that kind of stuff that can happen depending on what you're, what you're doing. Absolutely. So with the, with this one just sort of almost a side note kind of question. So you've been dealing with the, the world of, we'll call it the world of AI for, you know, six, seven years now. Have you seen, especially since it's the last, whatever, year or two years, it's really become a big thing and a lot of people are using it. It's very, almost a commodity to this point. In some cases, is that something that you saw? Were you sort of surprised that it, did it progress faster than you thought it was or was it one of those things that you felt like it was sort of a slow burn and it was sort of how you expected it to grow? Yeah, I mean, you know, a few years ago, AI was, was, was very advanced in terms of business. When you said, you know, we could do AI, the room fell silent and was hushed and was like, ooh, ah, let's see what you're talking about. And then years before that, when he said, hey, I'm a data analyst, the room was like, oh my gosh, here comes a data analyst, you know. But, you know, now you walk into the room, you say, hey, we do AI and they're like, so what, who cares? And, and now if you say we do data, they're like, whatever, you know, that's so what we write it back to, you know, so the days of the days of saying that, hey, we do AI, we do data, those are long gone. You have to be the next, the next tier year, your AI has to drive something and that has to drive something else. So your AI has to complement some aspect of the business and has to incorporate some other core piece of technology as well. So, for example, our AI incorporates, you know, telephony and the back ends of businesses. So that was that's what makes Barkey AI so unique. And then another core component of ours is that we make the the Barkey interface so user friendly. So people who do not have AI or AI experience can still use it and and and not be lost with it because a lot of times with people with AI, they can, first of all, they can tell it's AI and immediately identify that's a bot. But with us, you when you talk to it, it sounds like a real human person, pitch pace tone. It says om, er, ah, it stutters, it stumbles over words and there's pauses like rights happen there and you just feel immediately at home with it. So it's a it's a it's a great tool and we're really proud of the product. That's like, should be. It's a it's really cool and it's got a obviously got a really good application to it as well. I it's sort of near and dear to my heart in that sense, too, because I'm also one that I get really tired with the sitting on support calls or, you know, some call center somewhere waiting for somebody to get a hold of me and then them spending 30 minutes to go check something, you know, look something up, things like that. So it's the only way that that can be made smoother and and reduce hold times. And, you know, like you said, you've got a really good lack of lack of return calls. So it's basically get it resolved on the first time much more often. And that's that's huge. If I have to go call somebody back, I'm not a fan. So that's it's great to have a tool like that out there that can help you, you know, help you get it done the first time. Absolutely. Yeah, we just don't do that. Just I mean, we can incorporate that with emails, text messages, social media as well. So it's a versatile tool. Well, we just mainly since we developed the the text to speech and the speech to text component behind it as well, we just try to stick to a swim lane and stick with the call center aspect with it. Well, that's a actually sort of dovetail. So you've spent a lot of time in data analysis and, you know, now in AI and things like that. And one of the things that I think is when you get into that, one of the things I think is always a challenge is finding that swim lane is you look at the data and sometimes it's vast, especially large amounts of data. So you I think you tend to see large amounts of applications. So how do you how do you get into some of these situations and find where that that sweet spot is, is to look at all that, say, no, here's really where we want to here's what we want to focus on or here's where we can take this and and leverage it for our customer or our business. Well, for me, for instance, I mean, when we when we were doing this Barki AI thing, we had this application that was working really well for business to business. And then when chat GPT came out, we saw it and we instantly knew that right away that OpenAI came out with this great product. And at that time, the API was was private. But we knew that there's going to be venture capitalists with billions of dollars and there's going to be tech firms with billions of dollars. And they can and they can basically build a system like ours and have the marketing behind it to totally dismantle us, even though we had millions of dollars of capital rolling in. We just didn't have a substantial use of capital. A substantial user base to compete with billions of billions of dollars of business. And essentially, even though we had the the the runway going, we just knew that we couldn't compete with them because they had the dare to have the built in user base. They just had to say, hey, we got this new system is coming out. Here it is. It's going to chart, you know, two hundred fifty thousand dollars per seat, yada, yada, yada. And here's here's record profit record breaking profits. So we knew that we can do this. We had to find a different avenue of revenue. And how do we stay ahead of that curve? So we're looking at where where where is nobody looking at right now? Where is that one area where nobody's looking at? But everybody needs something. So that's where I went back to my my background when I have 20 years of calls in our calls in our experience and always thought of, you know, what if we had some what if we were would there ever be a call center, just a eye? Because we just got out of the pandemic. Everybody's working from home. You know, many of many call centers were, you know, trying to pay higher amounts of money to people. And I'm not trying to put people out of jobs or anything like that. I'm just trying to see how can we get people to pay less money for. Easier questions to be answered by a eye. So the so human workers could do what they do best and answer complex questions. Because, I mean, working from a call center, I worked at T-Mobile for a number of years. I would get bombarded with people asking the same question over and over. What's my balance? How many minutes back when they had minutes, a minute plan for T-Mobile, you know, 800 minutes for thirty nine ninety nine minutes. So I have left my plan. Well, let's this many minutes, you know, even though they have a self-service option. But I answered those questions over and over again. Wouldn't it be great if I had a level between me and them that answer those questions for them? And if they couldn't get answered, then I passed to me. I mean, that would save me so much money to save the company so much money. So that's what I was looking for. So when I proposed that to my developers, my developers looked into it with R&D. And I said, yeah, this is not a problem at all. Let's do it. So we did that partner with an art with a call center partner. We have known them for years, rolled it out and it was successful. And that actually, it's that's a great story, because I think it goes to something that I talk about a lot and have come across a lot with people is that you you're always going to be served well served to spend some time, I will say, in the trenches or something, because you because you spent time in the call centers and you understand things like you know, call center gets the same call over and over and over and over again. That there's those are those kinds of insights you only get because you're part of it. You're usually not going to get that just sort of looking at something from a, you know, from an ivory tower, 10,000 foot view of, oh, this industry looks like this. It's not till you get in and you you do the dirty work, as it were, you get down to the, you know, the the direct contact levels of, you know, with customers that you realize, oh, this is what this business is like. These are the kinds of problems that we see on a regular basis. And it opens up opportunities like that to say, hey, I guess I answer this. Yes, we've given them a website. But if we could have, you know, effectively my voice answer that because it's not a complicated problem, then that's that saves time. And that allows, as you say, the human to do the more complex things and not get distracted by all those little things that, you know, that could be automated or or quickly addressed. Absolutely. And that's one of some of my best managers I've ever had. We're the ones who are, as you said, weren't born in an ivory tower. They've done times in the trenches. I mean, as you put it, I've been on the phones. So I've I've been the one to my touch, my touch, your manager. I've been yelled at. I've been cussed at. I've been, you know, I've worked retail before, so I've been spat on and cussed out and all that stuff. So I've heard it firsthand. So I know what they have gone through and I don't want people to go through that. So I do everything I can so we can prevent that. So that's why that's why so unusual. That's why so unique to talk to Barkey, because Barkey has this unique personality to it that I can't quite explain it unless you actually have to go in and talk to it yourself. And that's why we have the app, the Barkey app, where you can actually talk to it and has a unique personality, where it's not judgmental or it's not cold. It's not heartless. It has a heart and it's warm and helpful. And you know, you just want to just have it be your friend. And I know I I'm biased by saying that, but, you know, I could show you emails of people who send me, you know, stories of how this has helped nurture their their relationship with their child who hasn't spoken in five to 10 years because of some kind of developmental disorder they've had. And Barkey has helped them open up because of the way Barkey is been programmed. So it's it's it's heartwarming to hear that. So that's the way we program Barkey is to be very caring and sympathetic to your needs. And we will pause there. We're going to come back next episode. Well, in a couple of episodes, we will come back with a part two. We're going to continue and really going to dig into some some value that AI brings and some of the things that Chris and his with his customers and his company, some of the things he's found that to me were very interesting AI applications and where people are finding uses for it. And so hopefully these will be one of those episodes or these pairs of episodes where you get a lot of good product ideas and service ideas. And how can you actually leverage some of these tools out there? Because that's really where we become better developers is where we're taking our skills, our experience, the tools that are out there and finding ways to either solve new problems or solve problems in a new way that is better than whatever it was we used in the past. That being said, I'm going to let you go out into your day of becoming a better developer and embrace it, whether you're using AI or not. Go out there and have yourself a great day, a great week, and we will talk to you next time. Thank you for listening to Building Better Developers, the Develop-a-Noor podcast. You can subscribe on Apple Podcasts, Stitcher, Amazon, anywhere that you can find podcasts. We are there. And remember, just a little bit of effort every day ends up adding into great momentum and great success. Please check out school.develop-a-noor.com. That is where we are starting to pour a lot of our content. We've taken the lessons, the things that we've learned, all of the things that make you a better developer. And we're putting it there. We have a range of courses from free short courses up to full paid boot camps. All of these include a number of things to help you get better, including templates, quick references, and other things that make us all better developers.