Eric Vaughan is the The Gen AI CEO 🤖👑 He has adopted an AI-First mindset at IgniteTech and GFI Software, that has fundamentally transformed the mentality and operations, even leading to significant changes in team composition. IgniteTech's new AI innovation, "My Personas," aims to create avatars of key organizational figures, enriching their knowledge base in real-time and making them more accessible without overburdening them with repetitive queries. Eric emphasizes the importance of developing an AI data structure and an AIPI (AI Programming Interface) to ensure clean, reliable data input, mitigating issues like AI hallucinations and enhancing overall AI performance. Thank you for joining us IMAGINE AI LIVE IMPACT New York at Cornell Tech hosted by Steve Metcalf Elizabeth (Liz) Ngonzi Chris Madden and interviewed by Chêrénique Chase
Transcript
Hello and welcome to the Imagine AI live podcast. I'm your host for this event share unique. We are here in New York City at the impact event and I have with me a guest today. Eric, can you please introduce yourself and tell us a little bit about what you're working on? You bet you bet thrilled to be here. My name is Eric Vaughn. I'm the CEO of Ignite tech and GFI software to software companies that are enterprise software and small and medium business software. That's why the two different two different markets, but we're all part of the. Same company and I have the great joy of leading these companies through this AI revolution that we're all going through right now. We're working on some amazing, amazing technology that's kind of game changing. But first we had to get our team. And for the last year, I've been working to transform our team fundamentally into what we refer to as AI first, meaning you think of AI first before you do anything else interesting. So how has that changed the day-to-day operations of the business? As you've been instituting that thought process in your employees, it's funny you say day-to-day operations, but I would say it's changed the mentality. Ohh. And that was fundamental. We actually had to, for the last five quarters now five consecutive quarters, we've been doing different things to bring the team that we had. Along to get them up to speed and energized and really believing in the AI and unfortunately that only work with some of them. So I'm going to have a talk later today at the at the conference. I'm going to show a chart that I'm not thrilled about. I don't love the results, but I have 23% of the same people that I had five quarters ago. Wow, that's not a reduction in force. That's a change of people. And that was after five quarters of investment. Training, education and the like. And we found that you can't make a lot of people change if they really don't want to. Absolutely. So can you describe what it is your companies offer a little bit? What is your ideal customer and how would they use your service? Tech is a an enterprise software company. We have over 40 products in our portfolio. Wow. We're a global company. We're 100% remote long before the pandemic. That's been our MO. We have products that run the gamut. Of different, you know different solutions, anything from marketing and sales intelligence to database software. We actually buy companies, we are an M&A based company and the software is used by very large multinational global companies, many names that people would recognize, you know all throughout the world. For the last year. Our key focus though is in developing AI technology and we've already shipped 14 AI new innovations for our enterprise software suite. And we're working on another one that was announced that Imagine AI in Las Vegas and will ship this quarter. And we're really excited about that. Awesome. Congratulations. Well, when it ships, you can congratulate you, but thank you. Yeah, fair enough. And it's called it's my personas. OK. So this is a different concept than a lot of other people are taking. We, we see a lot of avatars in AI right now. We kind of understand that it's kind of interesting. We're taking a different approach. The idea is one of cloning. The idea is one. What if you could take the people in your organization, the subject matter experts that you need constantly, the people in the company need head of HR, head of legal, head of product, head of customer service, and make them available in an avatar form. So it's people you know, you recognize them, you recognize their voice, and that avatar or that persona is embodied with all of their particular knowledge. And here's where the magic happens. So a person has an interaction with these. Personas asks them a question, let's say an HR related question about paid time, paid time off, the avatar finds or the persona. Actually finds that it is. It doesn't have that answer to the question and it can in real time if enabled, contact the human behind that persona using an app on their phone and actually interact and give them that question. The person can give the answer back, it comes back to the person and now it enriches the database with that answer. So the next time that question is answered, it will be available through the persona and the response to this. As we've shown it to customers over the last couple of months has been unusual. It's emotional. They don't care about the technology at all. They immediately start thinking, who would our first persona be? Who would, is that that we would want to replicate like that. And it has some other benefits like multilingual, you know, Yeah, because LM speak many languages, suddenly you're ahead of HR can speak many languages. Yeah, it's, it's really quite remarkable. Super excited about it. Just takes me down the path of so many questions. So once somebody's been replicated, I mean, are they still working for the company or are they pretty? I mean, are you just on permanent vacation now? How does that work for the employee then? That's such an interesting question that no one's ever asked. Yeah, right now, I don't think we've envisioned that it's replicating. Then goodbye. I don't think that's the idea. That wasn't the idea. It was literally to make them more available and more present, let them focus on more important things because let's face it, organizationally, we ping people all the time. Is an operative word. Ping me. Ping me, ping, ping me. Right. Yeah. And what if we could stop that noise? I mean, pings are great for submarines. They're not great for people, you know? So maybe maybe we can allow those subject matter experts to do more proactive and beneficial things for the organization rather than constantly answer the same questions over and over and over again because they haven't got around to documenting or the people don't look at the documentation that's out there or any number of reasons for that, Right? So, yeah. So then how do you program this? Rosanna, how do you get that information from the employee to their persona so that it's now able to communicate in the way that they would? Well, first they set up their their own personal presence. They design what they want their environment to look like. Video is recorded, so you get their image, their voices recorded. So it's trained. So that's all available. And then the way my personas is built, it can take data from any number of formats. It could be an existing database, it could be documents, it could be a. Google Drive full of documents, it's ingested In AI, there's this concept of something called RAG of retrieval, augmented generation. And so we're using RAG techniques for that. And we've got a methodology to bring that all together. But then it keeps getting supplemented by these answers with the actual human owner persona. So how much time approximately would someone need to spend to make their persona? Do you know or sure? I mean, I mean, this is a process that takes minutes to set up. The persona, the actual image and all the documents isn't ingestion by my personas, the software. It's just, you know, it's probably more time gathering. You know, if I came to you and said go get all of your important documents, you'd say, listen, I've got emails, I've got documents, I've got spreadsheets. But we can take all of that and it doesn't need to be consolidated, right? We can, we can train the AI to take that data and actually categorize it in a way and allow them to filter it. So yeah, it's a matter of hours to get it all set up and running. Amazing. So we're here at Cornell Tech. On a university campus, what would you say to people that are entering university and what they should be studying or going into with AI here in present? I just had a conversation with a young man who drove down from Rochester, NY. He's part of Cornell, but he came down here and we had this exact conversation. So specifically I'm, he said I'm a CS major and I'm focused on AI and ML and I, we talked about that. I said, you know, ML certainly has its place 100% been around decades. Absolutely still important. Don't let that be your focus. I can tell you that as we're interviewing and hiring AI innovation specialist and we've hired dozens in the last couple of months, when we hear ML, we're worried that they're what we call bottoms up AI people. And that's really not the, the approach that we want people to take where Jen AI has given us kind of a top down approach, right? And I also talked to him about something that we believe very, very importantly. You asked about the, the documents and, and the knowledge in my personas. We think that the real strength of what needs to happen is people need to build what is called an AI data structure. Because guess what? AI has the same problem that we've always had in tech where data is in many different formats. And so you start hearing the word hallucinate. And every time I hear that, I think, well, the poor AI, once you're not feed it garbage, what do you give it? Good data, Maybe something good can come out of it. A lot of times, hallucinations are the fault of the user, not the fault of the AI. It's just the AI is so eager to help, it tries, right? And so we, we believe in this concept. We think this is a fundamental principle of building what we call an API data structure and creating what is an AIP I not an API, an API API, an API for the AI to be able to feed it with information in a cleansing way. So we're very focused on building these AI data structures. And I talked to that young man, same same exact thing. So, yeah, fascinating. Yeah. Versus coding. Don't worry. About languages, that was the other thing I said to him, please do not worry about hey, I know Python And I know R and like stop with that because he knows all that. You know who knows that Yuri guy over here, Regal, whatever you want, right? Exactly. Yeah. Yeah. Well, we are coming to the end of our time even though I have so many more questions. But all I will leave off with one last question. What would you say to someone who is thinking about coming to the next Imagine AI live event? Unique part of imagine AI that I've really appreciated I think the term was coined for the first one in Las Vegas that it's the Woodstock of AI. I love that concept because Woodstock conjures this image of you know, cross section and people from all corners A lot of the other AI conferences right now are too heavy tech focused and I loved the opening Steve Metcalf in his opening asked, you know, how many people are technologists practicing how many people are you know in the industry and. Wanting to learn about using AI and that was more than the technologists and I think that's great. So anyone can come. You don't need to be an expert. And I love people that are walking around saying I'm here because I need to learn, but I don't know what I need to learn. That's exactly what I imagined I should be. Awesome. Well, thank you so much for joining us today and sharing your insights and your exciting project that's coming soon. Thanks so much. It's a pleasure to be here. You guys run a first rate event and it's really a pleasure. Thank you very much.To view or add a comment, sign in