Create an AI Clone of Yourself! | Rob LoCascio URAI | AI Outcomes Podcast Ep44

Yeah. So, previously starting UR AI, which I started two years ago, actually yesterday. I ran a company for 28 years called LivePerson, and I invented when you chat on websites for customer care. That was in 1997. So I invented this chat on web and then took the company public in 2000 and then grew to being a leading conversational AI company. About 3,000 employees. A little over half a billion in revenues, we were listed on Nasdaq for 22 years. And funny enough, in February of 2022, Fast Company put out the top 10 AI companies, and we were number one and OpenAI was number two.

Because we had a half a billion revenue at the time because we were running all the contact centers, the technology stack for all these large telcos and banks. And we had automated about 60% of a billion conversations at that time. And so, you know, I was deep into it, but I decided to leave three years ago to pursue things beyond the chatbot and enterprise and customer care. Because I had a different vision and I didn't feel like I could get there doing what I was doing, which was really SaaS enterprise selling.

That was my life. And I was like, this is just not gonna be the future. But I'm stuck in it, so I need to get out, so that's what I did. I broke out of the SaaS jail.

Yeah, I mean, first like my framework for looking at the world has always been a little contrarian. Like when voice technology in the contact center, that was it. You know, I was always about, I had this big slogan, kill voice, you know, and I wanted to get rid of voice and I talked about how it wasted a lot of our time and we were on hold and all that. So I took this very contrarian view to the contact center. And so when I looked at the world of AI and I've been so deep into it, I started to think of the alternative because what we've been fed is this perspective that there are large language models that have aggregated our knowledge, by the way, without our consent. Like it took our knowledge off the internet and messaging channels and podcasts and things like this, books, videos, and the concept is that there's this generally giant knowledge that is now owned by a few companies that own foundation models like OpenAI and Claude and Anthropic and others. And there'll be like giant data centers, which costs tens of billions of dollars to build. And then beyond that is now energy consumption, which is nuclear power and things we need for these GPUs to be in these large data centers. And that to me struck me as being tremendously non-entrepreneurial, which is that, and you've been through this, the revolution of technology is an entrepreneurial endeavor. It's small companies that have big visions that go and execute on those visions and bring them to life. In this world, it's a continuation of the people of Google's and Amazon's. And yes, there are the OpenAIs and there are the Anthropics. It didn't sit with me. It also didn't sit with me about the idea that our jobs are being taken and what it does is it's relegated, they've given us a perspective that we're not that important, that humanity and individualism of humans, it's kind of like not that important. Like I can make an AI salesperson and we can have an AI CEO, we can have an AI doctor and that these things are self-learning. So I felt that there's an alternative and I felt that it should start with humans and we as individuals like you. You, Ben, you're an individual's person who has such deep knowledge in selling that you're helping large enterprises sell, and they have access to AI. They can make an AI Ben, but they're paying you because you have this X factor. And so I felt like it should be humans first, and then the AI technology should support the human to be better, to get more leverage on your time, to be able to be your best self.

But not to just replicate you and throw you out in the dustbin. So I see a world where there's hundreds of millions individual AIs on a platform that we, these AIs are helping us that we're sharing ourselves. That we can monetize ourselves. And I think that's gonna open up a whole world of opportunity for us and not us losing our jobs. It's just, I don't see that world. And I think what we're seeing today is there is an uncertainty for most people around the safety of AI and how it will help us be better and create better community and be better companies. As technology companies, and I just think there needs to be an alternative. And I think that alternative is here and we're gonna see in the next few months the split between individual AIs and large language models and general purpose AIs.

Yeah, so I want to connect one dot because you brought something really important since we came from the contact center, which is, you saw this too as I was there when you too, where the internet came about and websites came about. And digital engagement came about and the concept around all of that. And it was true, which is that, look, we as a company are gonna put all the information about our company and our products and our services and everything you really need to know on a website. And that should reduce volume of people coming and calling us. And what we found, and you've seen this too, is like contact center volume kept going up. We never saw a contact center volume going down. And what that is, humans by nature have a sense of trust when they're interacting with each other. So even though I can read it on a website, and I may see it somewhere digitally, I have something, especially for high ticket items, banking, telco, healthcare. There's a need to just get that affirmation to have this relationship. So what the basis of that understanding what I wanted to build with UR AI was a way to truly replicate you as a person and you knew this too, like in contact center, there were these crazy great contact center reps who just loved talking to clients. And even in that world I see is that if we could replicate that human, their passion, their individuality. How they reason and then their voice, like truly build an AI voice that feels alive, that this could be, we could get humans in the game. So we did that. We started over a year ago and we built a thing called a human life model, and we able to understand and extract out what makes Ben, Ben and what makes Rob, Rob out of your personal stories. From those personal life stories, we then can build your knowledge of your business. We have things in which we take your frameworks of how you think, and we apply that into your AI. So all of that is part of the UR AI platform, which then allows you to give people access to yourself. Like, Ben, I can clone you now, and you could give your power of how to sell to millions of people. And you can't do that on a one-on-one basis. It's hard.

Sure. So the process is you come into the platform and actually the platform will be out for general availability in about four to six weeks. So we now have our beta tester group that's on it. So the process is basically you come into the platform, you record your voice, which takes about 30 seconds, and we are able to understand like your tonality and how you accentuate in certain ways. You may be excited at one point, or loving and caring when you talk. We understand that through your AI, we make an AI model of your voice. Second thing is we get your personal data sets and your professional data sets. You just upload them into the system. You could record them, in my case, I record every day. And I'll talk about the process of training your AI, but you put that in and then the marriage between that voice and your knowledge in what we call the human life model, allows you to come alive. And so, I'll give you an example. We do have a sales coach on the platform and one of the, it's a very seemingly mundane use case, but instead of going to OpenAI or Claude and saying, write me an email because I'm about this, I want to pitch someone in the enterprise, let's call it a CTO, and I'm gonna pitch them by email, there's a sense now that those emails seem generic or the person who's reading that email, who you're pitching may think, oh, they didn't write that. But I want to go to Ben's in this case, we have someone else. A few sales coaches where I want them to write the email. I want their AI to write the email for me so I get their view on how they would pitch this use case for me. I've got a CTO and a financial services company, and here's the background. How would you write the email for me? So it's a very base level thing. The next level up is help me write the presentation. So you could create the presentation now, because we have MCP services that connect to Google Docs and all that, you can write a document so your AI could write a document for me to help me present a sales pitch. Okay? We have doctors that do training. We have a doctor on the platform. We trains around concussions, other doctors. Those doctors upload patient information and that AI goes in and it looks at all the information and gives a perspective on that. So these AIs not only give advice, but they also can do action now. Okay? So they can drive action. We have accounting people on the platform who look at small companies and their finances and their profit and loss statements, and they can look at those and make assessments. And make changes on how they're billing or how they're taking in money or so forth and so on. So these AIs are not only there to give advice, but they're there to take action on behalf of a client who's paying you. And that's the power of these things.

That's correct. And they've spent the time to.

Yeah. Drag a file, you could drag a file of like your x-ray and your knee, and then their AI can make an assessment.

No. And if you go on to an individual level, like if I go into ChatGPT and say, let me give me an answer the way Ben would, or give me an answer the way Rob LoCascio, I have a lot of written work like you. I've been out there, and so a lot of people like I'm an entrepreneur, so people want entrepreneurial advice. If you go into ChatGPT and say, give me entrepreneurial advice like Rob LoCascio would, here's my situation. It biffs the whole thing. It's not me. It's not how I give an answer. But you would never, by the way, you never know that because you think, oh yeah, it's got all the knowledge of Rob, it's got all the knowledge of Ben. If I'm asking a sales, it does not. So we know that. So here's the big part, is that like small business owners are becoming really aware of the fact that every time they put more and more of their data into ChatGPT, ChatGPT seemingly gets smarter about their business. ChatGPT, for instance, is not self intelligence. It robs your knowledge from you. You willingly may give your documents to them and you unknowingly know that they may use that to train a model to be able to answer questions in your field of work. It's a very dangerous thing. And what these small business owners are starting to realize is like, I can't do this anymore. If I have an alternative, I can't. Because it will take your practice over. If you're an accountant, it's gonna take your practice. And so let's stop that. Let's have you take control of your data on our platform. You control all your data. You can delete it, you can own it. There's a delete button. They don't have a delete button. And so it's really important you delete and manage your data and then you can share that and monetize it. And that's the whole power of the platform.

So we have a model called the Human Life Model that we created. And so this protects your data and we do use large language models locally that don't share your data for training. The only place we use those is when we are allowing you to have world knowledge. For instance, when you're talking to your own AI, so it starts with you talking to your AI and using it like you would at ChatGPT, but it has all your context. So when you talk to it, you can say, I wanna take a vacation with my family. And it understands the context of vacations you maybe took in the past anyway. You use it like that. We connected so that your AI has world knowledge. Then you can make your AI better. You can, and that conversation gets saved into memory. So you can constantly through chatting with it, it's like a Tamagotchi. You make it better.

The second part is when you share it, when you share it, there is no world model on it. It is now only touching your dataset. In all cases, your data never bleeds or gets leaked into a public model for training. It only stays within our confines of our platform. So you don't have to worry about that.

And like we have Sonya, who you met, she's new here. She started about two months ago. And when she started, she actually was working on a marketing presentation about something on social media. And she had my AI look at it and review it before she even talked to me. And then she even asked it like she showed me, she made a video and she's like asked like, what do I need to do here to be successful? And the AI told her about what's important in this company. And so, yeah, I see the enterprise version of it to me is really about you'd be able to get the power of the employee shared between people, and that's what I see today is the power. I use it as my primary AI. Now, for instance, I do a full, I go on and I get a realtime feed of like, what's going on in AI in the world. I'll say, make a LinkedIn post. It has my tone and voice. It then makes a LinkedIn post, and then it goes ahead and posts it. So like I use it for things like that. I just use it to problem solve because it has my context. I also put personal information in it, but there's a private way in which even people who are engaging me can't get to. So like, I have my blood work up in there, so it has all my blood work and it understands my health. So I'm really using it. We're using it as our primary AI, and then from there I share pieces of it. So people can use it like Sonya. And then from there I'll monetize it one day on the platform around, I think, you know, around people wanting entrepreneurial advice so they can pay for it. So that's one part. By the way, the other thing I didn't, I forgot to mention, you can do video, you can do the, we have these little thoughts that you do images and video and people could ask you about like ideas. So you can abstract your data set into ideas in full video. So you can do an online webinar of you talking, explaining how to do sales training around a topic, and they can chat with you as your video's going on. So we allow multimodal in the platform.

Yeah.

Yeah. And look, I invented that chat window for customer care. I saw, I've been involved with chat since 1997 and I'm with you. I just wanna get rid of that thing, and it's very limiting. And one of the big limitations we found is that if you just put yourself out there on a chat and made yourself an AI chat. People don't know what's behind you. So like it's a lot of cognitive load for somebody to say, hi, how are you doing, Ben? Great. Oh, I'm trying to sell someone and can you help me sell Strike? Great. After a couple of turns, you're like, what do I ask this guy? And yet there's a lot more going on in your data set. And so what we did is we gave you a way to create these things called thoughts, and you can create a thought. And a thought is a small abstraction of something in your dataset. You put a visual with it, either a video or an image, and then people could read some text, look at the image, they could flip it around and actually chat with you about that thought. Just that thought. We can do these videos. And I think video is really, I mean, what's happening with the video models and we have them in our platform like Kling. These are, Sora. What they can do for you is they're getting so good where you can talk your way through a presentation. The part we're missing right now is a little bit like the actual visuals of presentation being high quality, right? And not losing text and things. But my ability to talk and even have you interact with me, is there. And so I think we're gonna see a lot more. I also think there's gonna be huge renaissance. We're working with writers right now, very well known writers, and even people who have been writing articles for like Vanity Fair and Esquire and who have great books out, but they never got option to make movies and, you know, it's too expensive. There's gonna be a renaissance on these people being able to create beautiful works of art, beautiful things that we can watch and engage with. And so I think there's gonna be a whole world now of UGC content, but a very high quality like movie quality, creative coming to market. And for you, Ben, too, it's like for all of us, like there's a movie in our head. There's a movie on how we sell, there's a movie on how we engage. We can create those movies. So that's where I see the power right now.

Yeah, I mean our, I feel like it was the one thing we really focused on, like imagery is one thing, but if you don't get the voice right, the images don't matter. So like we are very focused on the emotional aspects of voice, and we have real time. So you could talk to it like we're talking here. With very low latency. It's not perfect, but we definitely can capture your emotional range. And then there's some new stuff we're working on, I saw the other day, which is like, it's wild. Like I could ask you to sing like a song and know, like you can like in your voice sing out. Even if you off tone, you can like sing songs and you could laugh. There's so much we're doing there because it's a foundational level of our platform. If I close my eyes and we're just talking and I don't see you like on a voice call, I still have an emotional sense of you, and it has to start with the voice. Yeah, we, that's really important. It's part of the human life model is how do we capture you as a person, your stories, and then relate that to how you speak.

Yes.