Jimmy Huber: And welcome to another episode of the Info Tech Podcast. As always, I'm your host, Jimmy Huber. And if you'll remember, our podcast it always explores the intersection between tech and business. And today is no exception. I've got two guests today: the father-daughter duo, Nina, and AZ from Xeneograph. Guys, welcome to the podcast. AZ: Thank you. Glad to be here. Nina: Thank you so much for having us. Jimmy Huber: You bet. Yeah. I I enjoyed talking to you guys beforehand. I'll give you a second here to to share your background, but I think we had one other guest that that was a it was a family deal, right? I think the daughter was like the ops person and the the dad was kind of the technical, you know, mind. So it may be similar with you guys, but yeah, please tell us how, you know, you guys founded the company and what prompted you to work together. I think that's pretty neat. Nina: Yeah, absolutely. So funny that you mentioned ops and tech because that is the exact split of Xenia Graph. Yeah. Jimmy Huber: nice. Okay. Nina: I am the COO and co-founder, whereas AZ is our CTO and co-founder. So we really bridge business with technology with our two strengths and expertise areas. So Xenia Graph has been around at this point. Three and a half years we've we've been doing this and steadily growing. It's been honestly an amazing experience. Being early on in my career and getting to be an entrepreneur, I don't think I would have grown as much as I had in the last couple of years had it not been for this path that we took. And my background is more so in business and management. So I have my international international management for my undergrad and then for my MBA management. So I handle really customer service, partnership, sales, and marketing as well as operations. While AZ is our tech lead. And Xenia Graph, we are an IT consulting firm that really helps with data strategy for small to medium sized businesses. So that spans from data privacy, governance. We also do building agentic agents as well as automating business processes. So something that really differentiates us in this AI market that's booming right now is the use of knowledge graphs and how. to go more into that later on. But really what knowledge graphs are is they help AI and large language models hallucinate less and they make it more trustworthy. So we can go into the nitty-gritty of that later on but that's a little background on me and Xenia if you want to give yours AZ. AZ: Yeah. Hi everyone, my name is Aurelia Zovko. my first name Aurelia is very hard to pronounce, so I had my nice shortcut as A Z and I always joke I am A to Z. technologist, well over thirty years doing IT. very long time doing AI, 90s I built some expert systems and you know long journey around data AI and and bringing you know cognitive and help business making decisions. So And I think you know, like three and a half years ago L LMs was out and I I realized that's that's what I really like to do all my life and and working for enterprise corporation in a financial corporation, you know, you don't have that much time to do what you like and what is your passion. So that was the idea talking to Nina. I said I need help. I'm technical guy, I I'm that's it, I need some help and funny story she never won. wanted to do IT and anything around computers so I tried the Jimmy Huber: really? AZ: once more time and and I succeeded. So that's how we started. Nina: Yeah. Jimmy Huber: Wow, that's a great story. yeah, A Z, I think I'm probably the equivalent of you at InfoTech. If I didn't have my staff that was into the details and the ops and all that, it would be a disaster. So you you definitely need both roles. And yeah, I mean, I I don't have an MBA or anything like that, Nina, but you guys sound Nina: Mm. Jimmy Huber: like the real deal. So I'm I'm excited to get into your company because yeah, three and a half years ago, I don't think AI was near as mainstream as it is today, right? So you guys I think are are kind of on the cutting edge if you've been doing it that long. Sometimes, you know, my guests are startups and just just getting out there, right? So three and a half years, I think, in this space is is commendable. so you talked about hallucinations. I think a lot of us have seen that where, you know, even with ChatGPT or Claude, you're going back and forth and all of a sudden they just give you something that's completely wrong. Nina: Yeah. Jimmy Huber: I know I've done that before and I'll correct it and it'll immediately say, you know what? You're right. Well what what good is that, right? So I think that's an annoying problem that that everybody sees. so yeah, let's get into it. Wha so what does your product do to kind of minimize that, which I'm sure is a frustration for a lot of people? Nina: Yeah, absolutely. So the basis of large language models and AI programs are that The data that's looked at, it's looked at in a pattern. So when you add relationships to this data, when you have a knowledge base or a knowledge graph built out as the foundation, the data is then richer. It's it's more those relationships and context of the data make not just patterns that ChatGPT or Claude looks at, they look at the actual context of your specific business data. That's why we lead with being a consulting firm. Because each one of our clients, every use case, every solution is tailored to them specifically using their own data. So the the systems that we're building, the AI isn't just, you know, going out there on the web or or generalizing the data and looking at patterns. You know what you get back is going to be trustworthy and there's gonna be less and less hallucinations as the system learns. So I'm sure you've used ChatGPT for a couple of years, it knows you by now. and And I've even come into the interesting position where I'll ask ChatGPT a question, I'll ask Gemini a question, and then I'll ask Claude a question. And all three platforms give me a slightly different answer. So at the end of the day, which model do you trust the most? Where do you where do you really put all of your you know eggs in one basket, so to speak? So with the with the addition of knowledge graphs as that base of of knowledge, really, it's that we're making these systems look at the relationships and context. So an easy non-business way that I like to describe it, if you've ever gone to IKEA and bought a piece of furniture, there's a lot of pieces that come with that piece of furniture typically, without an instruction. Manual, you're gonna be sitting there for hours trying to piece together those pieces, and sometimes you'll be unsuccessful. So, what a knowledge graph is, it's the instruction manual, it's the instruction manual for AI and large language models. So we use them every day and don't even realize it. Google has the largest knowledge graph in existence today, and if you've ever used the Google map to search a restaurant or get the address, and you have or three different routes to get to to your destination, that is a knowledge graph working on the back end. So they're really widely used, but becoming more and more recognizable in the industry for how powerful they can really be with data. Jimmy Huber: Yeah, I love that answer. So you're basically training AZ: Yeah. Jimmy Huber: the the AI tools that you already have. So you guys aren't building out AI tool yourself. You're saying, hey, business, you know, you're a probably LLM agnostic, so you can just attach to pretty much anything, and you're helping enhance and probably shorten that learning curve to get it from launch to actually something useful. Am I on the fair way there at least? Nina: Yeah, i exactly. I'm sure AZ can give a little more background into the process. AZ: Yeah. Jimmy Huber: Yeah, for sure. AZ: Yeah, exactly true. And and I I'll throw another angle. me and Ina we were having a boot in a web summit, it's a big conference, a lot of startups and and and what's very interesting going around and seeing what people do and thinking in my head, I can probably do that in a couple of weeks. and and I think what's a challenge for small and medium sized businesses, right? They They need to be very fast and they need to re they need to use technology and they need to use AI. There is no way they can survive because someone can come around the corner, look their products, build it in a month and start selling, right? So, you know, I think that speed is important. So what we built to support that that small and mid-sized businesses, we build the accelerators, right? So kind of we have like maybe f twenty, thirty agents. All of that we talk about connecting data together, right? So we come to and help businesses saying this is my process, this is my data, and we can very quickly enable them to be more successful, right? Otherwise not be able to compete, right? You cannot manually go look things, you know, look data, spend hours, something that should be a kind of minutes, right? So that was our idea and and really core of our offerings to to help businesses doing that. Jimmy Huber: Yeah, that's great. So what's the response been? because I've I've got to believe that every business you go into is kind of on their the a different dot on the graph, right? they wanna grow, they probably want to use AI, they may or may not know, you know, how to do that, but they've started something, right? You guys come along then and basically blow it up, right? You I assume you you get some goals with them, you know, you see what's working, what's not, and you you're customizing every client, right? So that's that's a huge thing. What's the response been and what's like a a typical timeline for one of your engagements? Nina: Yeah. So I think with the industries that we serve, we serve a lot of regulated industries. So finance, insurance, wealth management, and trust is a very big component of working with these companies. So we we like to really plug and play with these accelerators. We're working in their environments, making everything tailored for them. So I think we've we've gotten on both ends, we've gotten people that are very wary of AI. I think There's a lot of fear in just the the regular landscape right now, not just only our clients, but fear that it'll take away jobs, that it might not do what it needs to do or what it's promised to do. So I think a lesson that we've also learned with speaking with a lot of these clients is AI might not be the best solution or AI might only solve a part of the solution and you might need to use different technology or processes to really make a long-term lasting impact on your business. So a lesson learned is not to take every project and since we do want to be a long-term trusted partner, really evaluate the pain points for the businesses and clients we serve, see how AI can plug in and make a business value within a couple of weeks. So that's really our goal. During our POCs or proof of concepts, proof of value, it's interchangeable. We really like to do that in a couple of weeks and show immediate results so that when we take on a full-blown enterprise project or work with our clients, they can see the metrics, analytics up front, and they know what to expect. So, you know, we've really gotten across the board. We've worked with smaller businesses, we've worked with enterprise level as well. So I think a very common thing is that a lot of them want to use AI, but using AI ethically and smartly is what will really differentiate them in the long run. Jimmy Huber: Yeah, I love that. A Z, anything to add to that? AZ: Nothing really, you know, like a common challenge is they're thinking about AI but always need to think about the data and you know, like whatever you feed to the model that's they will get back of course whatever training was there, right? So we tr like to say, right, your data is your gold, right? And you need to take care of your data. and and and we always try to say that's where we need to start look. You know, where is your data? Is your data clean, properly connected, semantically connected, how you use data, right? And then what Nina said, the key is the pick right use case because some use cases are not good, most likely will fail, or it will not bring enough value for business, right? So picking the right use case, automate some business processes and bringing value and be very fast and and adaptable, that's kind of key for us to and for business to succeed. Jimmy Huber: Yeah. Yeah, we had at a recording earlier today and the guests talked exactly like that. It's all about the use case. There's no one size fits all and there's a need for the business analytics side before you I think introduce the AI tools. Cause if you don't have a process or an o thing to automate already, AI can almost be a detriment, right? So are you guys business consultants as well as bringing together, you know, the the graphing and the the tech side? Or what what is your, you know Do you have to have some of those prerequisites in place to work well with you guys? Nina: Yeah, so we're definitely there to to be a resource for our clients. A Z has over 35 plus years, as he mentioned, really building out these the production level enterprise using AI and the different tools that we mentioned. So yeah, I I I think that is really, really important in the sense where we work with different clients. every client has a different need, a different use case. So, like you mentioned, it can't be a black box. I think even if you're using LinkedIn, to give a sales example, LinkedIn or Apollo or a Crunch Base. Those are very good tools, but they're also more generic. So if you need a specific process, sales process, you can't just go and and change the criteria or change the buttons that that you have allotted to you. So I think that's where we bring the value where it's not just the technology. We're really there to serve as a strategist for your business. we bring in that data strategy along with the technology that we use, especially with AZ's expertise over over his career. It really it's a a big differentiator as well. Jimmy Huber: Yeah, I love it. Go ahead. AZ: And and just to add from yeah, just from the business domain very important, right? And and so via finance, wealth management, insurance, right? And we did you know kind of data IT strategies, architecture strategies, but but but the good news is that the the business domain is important, but you know, these these things are very it can apply to any any kind of business domain. So we just another client you know, for for supply and chain and and we did some kind of work with them, it's working very well, right? Even I'm not expert in supply and chain and we have another, you know, kind of more like pharma slash, you know, wealth and helping, you know, people cope the problems. And we had that solution and and worked well, right? So yeah, b business is domain is important, but all these technology accelerators it's very applicable to any domain. Jimmy Huber: Yeah. Yeah, that's great. So everyone wants to know about security as well. what is you guys' approach to, you know, 'cause you can accelerate, you can you can do all the knowledge, you can train it, but if those security parameters aren't in place, I can think that could be a really bad day for people that you know, for identity th IP theft, you know, the just the data leakage, there's all kinds of things out there that I think you can't talk of AI in one hand and without security in the other. So what's you guys' approach to, you know, combat the fear like you mentioned, Nina, but also to still give, you know, businesses the advantage to to go quickly and to s to scale but still be secure? Nina: Absolutely. So our our first and foremost, you know, goal is to make sure we're a trustworthy vendor. So, especially in the industries you we work with, like you said, there's a lot of personal information, information that you don't want getting out. So I know some of my friends even recently had their data breached from using specific websites or going to specific places to to get their blood basically their DNA information was leaked, so that's horrible. yeah. Jimmy Huber: wow. Okay. Nina: Horrible. So my friends are going through that process right now. And I think it's becoming more and more unfortunately common where there's these data leaks. I'll get emails for class action suits. It's it's an interesting time, definitely. So we combat this by plugging and playing into our clients' environments. So your information never leaves your systems, your information never leaves your environments. We are there just to exactly be an accelerator and and streamline your data strategy. strategy, data processes, whatever the use case may be. So you know we're we're there to to make things easier, not give you another headache or another AI platform that we just leave to you and and you know say bye after a couple of months. We're there for maintenance, QA and and anything really that our clients need from inception all the way to to after support. AZ: And just to be a little bit technical, we have something we call governance framework, right? Let's say you ask question and before getting answers, right, we we go through garnessing, can you answer this? and that part of that will be, you know, is the any PII data anything you should not answer and you should not set. On top of that we have another complete hallucination framework where we'll say, Okay, this is the answer. Is it really relevant? Is it based on the context, your context and then you know if we are not there then we don't we say we we cannot answer you, right? So and then on the data privacy we building complete software as a service solution around data privacy meaning can I use data, how I can use data, you know, how to protect the data, you know what is my compliance, can I use the data to train the model? Can I use data to move it here and there? And then when you think about fifty states in US and you're operating internationally, then GDPR is coming in and can I send data there? What I can do with data, all that complexity, we we have solutions for that too. and what we like to say we do privacy by design, meaning when you start doing something, you know, design privacy and data privacy is a key at the beginning, right? Not like I have everything for you now saying, Hey, let me now check, you know, what's going with your data privacy and and everything else. So that's kind of approaches we are using. Jimmy Huber: Sure. Yeah, and I think that's great. The the fact that it has built-in checks, you know, to check itself. You're not completely relying on the user. there are some guardrails in there, which I think is really prudent. so how do you train your model then when I I assume a common question is gonna be, hey, how do I know that what you're learning with client A, you're not gonna use for client B, C and D without exchanging, you know, data? not to get, you know, super technical A Z, maybe that's the one in your wheelhouse, but you wanna be able to use I would assume what you're learning, you know, with all your clients, but you still have to keep those guardrails up. So how exactly do do you guys approach that? AZ: Yeah, it it's kind of very straightforward r because we use the internal data, so right we when we go to LLM we say, Okay, this is your context, this is data and use this data to answer, right? So you know without that you can say hey I have an article five years ago and I can use that to answer, right? No, we are really tailoring saying this is the question and this is data you need to use to answer this question and this is your data, your approved data. And then while we're talking about data, you you clean the data, you ensure that you know you provide right data based on the question of what you want to do. so that's very safeguarded. And and for the clients, we sometimes approach they they run the LLMs internally, right? So if you do API to Chad GPT, you say okay, Chat GPT will not use my data. I trust them, but if you really want to be sure, right, you use your local models, meaning You run your models, everything in your environment, so nothing in theory should go out and and then that's where you protect your your you know, your IP. Jimmy Huber: Got it. So what's next for you guys then? you know, you're in a really fast-paced industry. You know, what's the next six months look like? Are you guys rolling out, you know, more features or where do you see, I guess, the industry taking you? because I I assume we're still in the early stages, even three and a half years in, I can't imagine what the next three years are gonna be like. So put the the visionary hat on maybe for a minute and and paint the picture for us. Nina: Yeah, absolutely. So we're always looking to grow, try new things with AI. I think it's a very innovative space to be in 'cause we don't know what's possible or what's fully possible yet. So it's really cool for lack of a better word, it is cool to be able to Jimmy Huber: yeah. Nina: see what AI can do. But I think specifically for Xenia Graph, we are, like we mentioned, IT consulting, but we're looking to get into software development as well. So in the next couple of quarters, next couple of months, we are rolling out with our first insurance SaaS model to serve the insurance industry. And then next year, first to second quarter We are rolling out with a data privacy SaaS model as well. So we're moving a little bit more into the subscription-based model. So that'll be interesting to see the difference between just not only clients, our processes, and how we're able to really merge consulting and SaaS models together. Jimmy Huber: That's legit. Yeah. As as someone that relies on monthly recurring revenue and our managed service model, I can tell you, yeah, it is a little bit easier than than having to get projects and stack them. But I think they go really well together, right? So I have to believe that's a great strategy because you're already building trust with a lot of your clients. It's a natural next step to be like, hey, here are some solutions that you can do ongoing to enhance and even get better. AZ: I just Jimmy Huber: I think that's a that's a great strategy for mid and the long term. so what what are some struggles that you guys maybe have seen from the business side here? I mean you're you're not a startup, you're definitely, you know, growing and expanding, but you've gotta be able to be nimble, I think, and and move with the punches. Does any story come to mind or example of of where, you know, you may have hit a ceiling or whatever or when able to, you know, break through? Nina: Yeah, so I think a lot of it comes down to just learning from your mistakes because we've made plenty of them. so I think early on we we were focused on technology, technology, and a lot of like you mentioned, a lot of people just aren't familiar even with the AZ: Mm. Nina: words or or the concept of it, especially with knowledge graphs not being as mainstream as AI. So leading with business value, leading with business lingo, more So, in the first couple of conversations, and then merging that with how that technology can accomplish. That's been a learning curve, and we've refined our pitch, we've refined our demos and how we do client outreach with that in mind of being business focused and then AZ: Yeah. Nina: bringing technology in later on because it is so cool what the technology can do. But if you don't present the value, then how are you know potential clients? really gonna know what they're investing in, what who they're working with. and I think that also adds on to just not taking on every use case and every client. I think a lot of you know AI companies and just smaller companies they're fast to grow, but that also gives you more bottleneck issues, more operational inefficiencies and and you're scrambling for resources. So I think as a growing business, my my advice would be it's okay to grow s grow s slowly but steadily to really, you know, outlast those first couple of years. and always be willing to learn and be agile. I think that me you know from my background how I take in my my professional career, my entrepreneurship journey is that it's okay to fail as long as you keep on learning and especially learning from those mistakes. so I think those are my side personally as well as Xenograph goes hand in hand. We we've learned a lot from the last couple of years. so yeah that would be I know AZ I'm sure has a different response to that. AZ: Yeah, I think of course I will get to technology so much changing and and I think our lives are changing and we want it or not, right? And our lives probably in three to four years will be completely different what is right now. So just keep a technology and life just figuring out h how to prepare when next new thing is coming and will come very soon, right? And and like models when I remember like three years ago, two and a half years ago. what you could do with models and what you could do now, it's completely different, right? So really adjusting Xenia to to be on cutting edge and and challenge is really to just read articles, understand what's going on and then pick what is really you think is value and and try and learn and and see if it applicable for you. Right. So that's kind of a daily or weekly struggle and and and and how to reflect all of that in our offerings. Jimmy Huber: Yeah. Boy, that was like a master class in in small business growth. Go ahead, Nina. Nina: Yeah, I'm so sorry. I was also gonna add we we've also taken sales and marketing courses because you know, y you can't cover everything in a business. You have to put on a lot of different hats. So find mentors, find people in your corner that you can rely on and and be a resource. 'cause with that sales course that we just completed actually recently, we were able to see there's such a different process with cold outreach versus warm outreach. So having really Jimmy Huber: Tell me about it. Nina: yeah, it's it's so different. So having a set plan, rehearsing and just trying it again and again, that's okay. That's how you learn. So that's just a quick addition. Jimmy Huber: yeah. Well you you kind of led into my next question because I tell my team, fail fast. That's the way we learn. we also did tech jargon early in my career and no one cared. the people that are decision makers generally aren't technical. They just need to know what it does for them, the value, you know, and all those things. So I I feel a a a kindred spirit with you guys because we're we're at the point in our careers where we're trying to figure some stuff out, but we've we've learned a couple lessons. So what you kind of alluded to this already, but for people that are maybe just starting out. And haven't got that far in, right? I feel like if you're you know graduating college or entering the workforce right now, you probably have some framework in AI or else you're kind of behind, right? So we'll give that to them. But there's so much opportunity, it feels like that you can kind of be a little scattered, right? I think there's wanting to grow, but then there's wanting to scale, and scaling requires simplicity, right? So what would be your advice? I mean, you got two really good perspectives here. I would love to know. You know, maybe the top one or two things that you would tell someone who's who's in that spot and just a little bit further behind. Nina: Say again, this is me. follow something that merges your passion with what you're good at. You're never gonna like everything in your role, in your job, but that is part of learning. So if you merge what you're good at with what you're passionate about and find the intersection, I think any obstacle, any any curve of I don't want to do this paperwork or I don't want to read this contract, it becomes part of something that you're working towards. towards a bigger goal and that'll give you that motivation to to really keep going and and at the end of the day, if you're doing something you love, it's not work. Jimmy Huber: Love that. A Z? AZ: Yeah, I I think for me is you need to be practical. I immigrate to to US and and all that story, right? So life is unpredictable. so for me you need to always think you know, I want to do something and I need to make a living and then I like to do something what I like. So I think that balance, you know, it's hard to to do, but I always thinking that You know, job is job and job could be boring. and and and if you can like what you're doing is best, but you need to really pick something that you you can support you and your family. and then if you have really passion, it's not through job, then that's you know, this is your passion and you do when you have time. and and of course learn technology, right? You'll not be able to survive without technology. you doesn't need to be developer, you don't need to create a technology, right? But how to use it, understanding technology, how they're working, it's a key because that will be kind of machines and humans will be much more, you know, together doing things right. So you need to learn that in order to really be able to survive. Jimmy Huber: Yeah. No, I think you guys are completely right. Hit the nail on the head all around. I would think that figuring out what you want to do and your passion and if you can actually form a job and make money at it is much easier these days. Or it seems like you can get there faster with the tools that are available. I know I've got kids that are, you know, in high school and and starting to think about those things and it's like, hey Yeah, you you have to know what you're doing a little bit. You can't just put everything in the chat GPT because it's gonna spit out something and you gotta use your brain, you know, as well. So I love what you Nina: Yeah. Jimmy Huber: said, AZ, that you know it's it's humans and tech. It's not either or. I think navigating those things is is is super important. so I really appreciate you guys' time today. This has been a fantastic conversation. if people, you know, I a lot of my audience, I think they resonate and they want to connect with our our guests. So What's the best way to get in touch with you guys or or Xenniograph in general? Nina: Yeah, absolutely. So you can find us on YouTube if you search Xenia Graph, as well as LinkedIn. We have our company profile and then we have our website, xeniagraph.ai. So any of those, please feel free to reach out. We're looking forward to to having a conversation and helping you out. Jimmy Huber: Yeah, you bet. We'll definitely put those in the show notes. so it'll be easy for people to link and and and find that there. but yeah, thank you guys both so much. It's been a great conversation and hope to have you back maybe in our next season, because I'm sure things are moving and shaking all the time in your world. so we'd love to follow along for sure. Nina: Thank you. AZ: Absolutely. Thank you. Nina: Thank you.