Latest / The Tech Matchmaker / Solving the Tech Stack Overload | Aaron Symbolik on Krista.ai’s AI Platform | The Tech Matchmaker
Transcript
- Gina Gibson: Welcome to the Tech Matchmaker, the show where contact center challenges meet their perfect tech match. I'm your host, Gina Gibson, and in just 15 minutes, we serve up the smartest, coolest innovation shaking up the contact center industry. Whether you're searching for your next solution or just love staying ahead of the curve, we've got the insights to keep you in the know. Let's dive in and discover what's next. I promise you it's going to be fun. Thanks for hanging out with us on the Tech Matchmaker. I hope today's spotlight sparked some fresh ideas and real solutions for your contact center. Want to dive deeper into the tech we featured? Hit up the show notes for all the links and extras and be sure to subscribe so you never miss an episode. And if there's a challenge you want us to tackle, don't be shy, reach out. I'll catch you soon with more innovation, more use cases and more ways to keep your business moving forward. Good afternoon, everyone. I'm really excited to introduce you today to Aaron Symbolic from Krista.ai. Welcome to the podcast, Aaron. Aaron Symbolik: Awesome, thank you so much Gina for having me, I appreciate it. Gina Gibson: Absolutely. Now, Erin, in the research, they told me 26 years in tech. Paychecks, paylossity, you've been a CRO, a sales leader of the year, and now you're in AI. So tell us a little bit about yourself, stuff that we wouldn't find on LinkedIn. Stay curious, stay bold, and keep exploring what's next. Aaron Symbolik: Something you wouldn't find on LinkedIn. I grew up in the tech space. Even as a kid, I loved computers. I loved the outdoors, but I also loved computers more. My favorite movie was War Games. I kind of hooked on technology then. So I think just I've always professionally didn't know I was going to be in technology, but that's probably something you wouldn't find on LinkedIn that I love I love technology. So I'm so happy that AI is where it is today. So yeah. Gina Gibson: Excellent. So I know with Krista, I usually say, give us your elevator pitch. But I think with Krista, you know, I'm a huge fan of the product. It's a little bit more than an elevator pitch. So give us the longest escalator ride of our life pitch. ⁓ Aaron Symbolik: I'll do my best. And I feel like I'm a lucky passenger here at Krista because our founder, John Michelson, he's a brilliant guy. I would put him as one of the top guys when you look for artificial intelligence. He came from cybersecurity and he was a serial entrepreneur. He started many software companies, but this is what I would call his opus or a swan song, if you will. And I'll make it really simple. He like set out to solve two problems and that it's the problem that companies have today. It's the one they're about to have the problem they have today. is that we've overloaded and overwhelmed our employees with the tech stack. I think on average there are 70 different UIs essentially we're expecting them to extrapolate all the value out of the software. And John would say, said, as a young software developer, I would develop the software and we get so mad that people just wouldn't understand it. And when he came to his senses, he realized we're the problem, the developers, because we're over developing software and we're trying to extract value. I a white paper, it was interesting. It was a Microsoft white paper and said, what percentage of Excel's capabilities do you think any one person knows how to use? What do you think that was that the white paper said? Gina Gibson: You know, I remember you and I talking about this and was it 9 % or 13 % something like that. It was just that really low. Aaron Symbolik: It was actually much lower, it was like three to 5%. And it's like now extrapolated across company's entire tech stack. We're expecting our employees to go in and take a mouse point and click and be data entry people. And so I think where the real passion comes in, this is what I really loved about coming on board with Krista, is that we can start to give purpose back to people's jobs. It's not AI is going to take people's jobs. it's going to be AI allowing people to move at machine speed and take the tasks away that people hate about their job anyways. That's the cool thing about AI is everyone's so afraid. when we're talking about change management, because we're really ⁓ at we're an enterprise wide platform orchestration platform. And it's a lot to handful to say, but enterprise wide AI and orchestration platform basically means that we're crossing the entire enterprise. You think about all their different departments. we can find multiple use cases and move at a very high velocity. So John's first problem he wanted to solve is let's use AI as that single pane of glass to extract the value from the software where we don't expect people to have to log in all the different applications. We can go through one single point and extract value by letting Krista go in. So let's let software understand people versus us trying to build software and make it understandable to people. Very different narrative. Now the problem that we're about to have is that Companies are starting to look and shop for AI the way they bought software. And who, why would we blame them? Because that's all they know, Gina. you know, it's like, so eventually you're gonna have 10, 20, 30 agents, chat bots running around your organization, you know, and they're still going to have that problem of being siloed. It's not working across the spectrum of an entire company. We focus on looking for outcomes, outcome-based solutions, rather than just picking off a task here, a task there. That's really the big element. Gina Gibson: You know, I love what you're saying. And it's a common theme I'm talking with more people all the time about AI is that really focusing on the employee experience versus, you AI is not taking the job. We're just going to make their experience at work a better one to have. Right. And I think that's a really takeaway for that. I think you've kind of answered this in being enterprise wide and not a point solution, but is there anything else, Erin, that you'd like to kind of Aaron Symbolik: Exactly. Gina Gibson: touch on that sets you apart from your competitors because I don't even know if there's really anybody else out there that's doing what you're doing right now. Aaron Symbolik: They're not doing this for a lot of reasons. And it's again, no fault of probably companies, frankly, the last couple of years, the of the LLMs like ChatGPT and some of these tools, if you think about the ChatGPT experience, everybody that first used it was like, this is amazing. You didn't need training on how to do it. So it's really opened up the possibilities of AI. So what a lot of companies have done, have found niches to get a solution. Now some of that's a money grab. And there are some that are really good solutions. So not bashing that at all because there are some bespoke solutions out that are great. But I'm old enough you can see the gray in my beard that, you know, back in in, you know, the 22,000s of the dot com era, there was a lot of companies that went under like, and that's going to probably happen here. So I think companies are a little afraid of what strategy do we go with? And the way we built Krista, we started building Krista about seven years ago. And John, being the serial entrepreneur, was very successful. In fact, 80 % of Global 2000 used at least one of John's applications today. He's been in this AI space since 1992. So this isn't his first rodeo. He didn't just go out and say, AI is the hot new thing to get into. But what he did is he went to his investors and said, I'm going to build something that no one else is building, probably doesn't want to build. And by the way, I'm not even going to give you a minimum viable product for six years. no venture capital, no investor wants to hear that, that we're not even gonna get a product for six years. So they just let, trusting John, let him just build Krista for the last four six years. And now we're at a point now where the differentiator and the reason why it takes so long to build is we a no code platform. Well, what does that mean? That means if a company comes to me and says, we have this very specific use case we wanna do, we don't need data scientists. Krista herself is the data scientist. We use natural language just like we do in those LLMs like ChaiTBD. tell it what we want it to actually do ⁓ we build these intelligent conversations we call them to actually produce the outcome. And I can certainly get into some examples. So the difference is if looking for a point solution, well, maybe I go find, you know, four or five, six companies that can do that. But why wouldn't I just go to Krista that can do that? Because that's table stakes. But then we're going to probably have two, three, four, five. Gina Gibson: Mm-hmm. Aaron Symbolik: new use cases because AI is moving so rapidly, I don't have to go shop around and hope that those companies are still in business when I know that Acrista that's well funded has been around and very unique in the marketplace. So, you know, from our competitors that we could bring something very unique. We can move people very at high So they're not looking to bring one use case for the entire year and have this big, huge, immense lift. and then taxi their IT department or outsource that. bring Kristin, we're putting up use cases in four to nine weeks and they're on to their number two, number three, number four use cases. They potentially put up six to 12 use cases in one year alone with ease. Gina Gibson: wonderful. And you alluded to it, you said you could give me examples. Share with us one of your top success kind of use cases that your user stories that you've put into for one of your customers. Aaron Symbolik: Absolutely, I can go on and on. I'll give you a couple really good ones that really show you the depth and breadth of what you can do. And we further achieve because she's Krista. but it's so a couple different. So an example of a client of mine called Guardian Alarm, they have about 600 employees, but they're an alarm company and they have their customer service oriented, right? They want to make sure one, they want to take care of their clients, they want to get resolution on there and they want happy clients, right? They don't think they're worried about churn. The second thing is they want to empower their agents to be able to provide good customer service. And so they have that knowledge there, but they can get to it quickly. And then that allows them more capacity to grow, not have that potential for churn or bad customer support, right? ⁓ they could kind of prevent that. So what they did is we created an agent within their Microsoft CRM, their iframe. where it sits on top of each client. So when the client calls in, it recognizes the client, it recognizes all the equipment. So instead of having to look up SOPs, FAQs for all that different equipment, Krista already knows the type of equipment, the model number, the make. So I can just start tapping away just like I did in ChagGBT, but remember it's trained on their data and LLM is not. LLM is trained on global data. always says that half the AI you'll ever need, you can buy off the shelf, but the other half, has to be trained on a company's information. So I couldn't go to another alarm company and actually do the exact same thing. I have to create a custom to them. And that was the one case. And I'll give you another one. It's a really cool one. For all the folks out there that may be familiar with the payroll industry, like Paychex, ADP, Paylocity, that was a company I worked for for many years. 7,500 employees, publicly traded company. Every company has state unemployment when they have employees. And that changes year to year. Gina Gibson: Yeah. Aaron Symbolik: and it could go up or down depending on how many unemployment claims they have. There's roughly 200 state unemployment forms that are PDFs that come every year that are I would call unstructured data. It's just an image, right, a PDF. So you can imagine a person would normally get that, read it, send it to the tax department, open a ticket in Salesforce, use math to figure out what the new SUI rate is, update the record, and then send that off to the client saying it's been updated. On average, it was five to 10 minutes. Gina Gibson: Right. Aaron Symbolik: Paylaw was getting 400,000 of these a year. Within six weeks, we were at 95 % accuracy. Now we're upwards of 99 % accuracy where Krista is handling all those PDFs coming inbound, reading them, opening tickets, closing the ticket, responding to the customer, and using what we call regressor model, AI modeling form that we use to figure out what the math is. We trained her on those 200 different types of SUI docs that come in. So when that information looks at it, she knows it's not like traditional OCR. OCR were documents where you have a system that says go up two inches go over three inches. There's the PO number. There's the account number. You don't need to do that. It doesn't matter wherever that PO number sitting on that PDF document. Krista can find it. Gina Gibson: Yeah. That's incredible. Is there any, mean, those are two amazing examples, but is there anywhere, do you have a customer that used the product in a way that you guys were like, ⁓ wow, never thought of that. Like that kind of caught you off guard as far as like, that's really creative or a different use. Aaron Symbolik: We get it all the time. So remember, because we're enterprise wide, we're industry agnostic, which is can be challenging, right? Because, you there are point solutions and they say, well, these guys do this one thing over and over again. But most of the time, those are table stakes and they're not hard to do. So it's very repeatable. Probably the one interesting one I got was ⁓ a title of ours and very successful. They spent months trying to build it out to where all these Gina Gibson: Yeah. Aaron Symbolik: unstructured emails would come inbound and they're trying to get that data and create these forms. You know, if you ever closed on a home, have all these different forms like warranty deeds. They had like nine different forms. Krista was pulling from multiple ways. They said, we want you to pull from email. We have documents. We want you to populate these, but then we might take a step further and we want to actually commercialize it to other title agencies that don't have lawyers on staff. And we want to charge them for this form. and actually automate it. So they have this one giant inbox where all the information comes in for each client. Krista ingests that, creates the actual form, and then creates an invoice for that particular title agency. And now they turn it into a revenue machine for their company. And that was very unique. We didn't think we could do that. We didn't know we could do Gina Gibson: That's very creative, creative. As we wrap, Erin, I've always been a fan of James Lipton. And so we end with a few rapid fire questions. So what is your favorite word? Aaron Symbolik: Yes. My favorite word. that you're catching me off guard here. My favorite word would probably be, would say inspirational. I don't know. Gina Gibson: Excellent. What's your least favorite word? Aaron Symbolik: Can't. Very simple, can't. Gina Gibson: What excites you the most about life? I love it. I love it. What excites you the most about life? Aaron Symbolik: What excites me about life? family. I four kids, two are in college. They're all unique and different. I've been married for 23 years. I was in the Marine Corps, but I live in California. I love the weather. I like mountain biking. I love taking walks with my wife every day. My family is what drives me, ultimately. Gina Gibson: Fantastic. Erin, somebody wants to learn more about Krista.ai, can they reach out? Aaron Symbolik: I think they should reach out to Gina. I think they should reach out to you. forward. It's who they should reach out to. Yeah. Gina Gibson: And you. Excellent. Erin, can't thank you enough and I can't wait to have you back when you guys do the next incredible use case to share more with us. But thank you so much for your time today ⁓ and a great day. Aaron Symbolik: Yep. You too, Gina. Take care. Thanks for the opportunity. All right, bye. Gina Gibson: Thanks.