Motley Fool Hidden Gems Investing - AI is Much More than ChatGPT

Episode Date: May 21, 2023

An artificial intelligence program could soon create a digital version of yourself. So, what are the implications when that’s possible? Rob LoCascio is the founder and CEO of LivePerson, a company that develops conversational commerce and AI software. Ricky Mulvey caught up with LoCascio to discuss: - How artificial intelligence will change how you interact with brands - Starboard Value’s campaign to change LivePerson’s board and CEO - The unlocks that quantum computing could provide tech companies - What natural language processing programs are revealing about interspecies communication Companies mentioned: LPSN, GOOG, GOOGL Host: Ricky Mulvey Guest: Rob LoCascio Engineer: Rick Engdahl Production Assist: Dana Corl Learn more about your ad choices. Visit megaphone.fm/adchoices

I call it like it'd be your pet, you know, like your pet dog, you know, like something you love and cares for you. And then I think the brands have to have their own AIs, and I think these AIs connect. And so that's where we look at it, we call it the digital twin, is that ultimately the brands have a digital twin of you as a representation of how you engage with their brand. And when your digital twin shows up, it can connect with theirs, and they will get things done for you. I'm Mary Long, and that's Rob Lacascio, founder and CEO of LivePerson, a tech company that uses artificial intelligence to help brands talk to their customers. If you've been on a web chat with a customer service bot, you've probably used its technology. Ricky Mulvey caught up with Lacascio to talk about why your interactions with companies could soon change drastically.

How Lacascio is navigating a stock price drop of more than 90% while also facing an activist campaign threatening his job and the ways natural language processing could transform how we communicate with, wait for it, animals. It's going to get pretty weird, but it is a very fun ride. Right now, going into 2023, what are some of the use cases for live person's products? Where could I specifically have encountered it? Sure. So we provide all the messaging AI platforms for some of the largest brands in the world. So if you're on Delta Airlines or Verizon or United Healthcare or Home Depot and you're messaging or chatting, that's our platforms. I was going to ask where you fit in the competitive landscape with call centers

and other AI chat products, maybe even companies like Twilio. Yeah. So we were the first out seven years ago. We pioneered the messaging platform. And my goal back then, still holds today, is that I want to get rid of the traditional 800 number. I want to kill, I used to kill the 800 number. I just think this idea of talking by voice, pressing one, two, three, and what we call an IVR should be gone. And that's what I've been after for the last seven years. Messaging went a long way. We added a lot of AI into it, but now with generative AI, like there's a path now to kill it once and for all. Well, and part of that path for you involves this new product where voice is back into calling and human speaking with essentially AI chatbots

and getting real-time reaction. Obviously, all the generative AI stuff makes the machine understand natural language better than ever before. The other part of that is the data set that's going to power that. And the next part is the actions. I want to buy something and there's got to be actions off of that, which generative AI itself doesn't quite do that. And that's where we also provide actions so you can buy and get service and support. And with your new products, it seems that almost the 800 number might be back a little bit where people, humans are able to interact real time over voice with AI chatbots for customer service applications. Yeah, we just launched our new voice AI and platform. And that's really what I believe

will be the final attack on the thing I've wanted to get rid of, which is traditional voice where you're sitting on hold and you're talking to a human, we are now offering voice AI to use large language models and generative AI to power those conversations. So it's really fascinating. Like what we're finding is that like, take our, we looked at our top airlines, people ask the craziest questions. Like people were asking, can I bring an iguana on a plane? And in the past, you can never make a traditional bot that would handle iguana. You would never even think about making that. But with the large language models, it just picks it up because of the data set. And that's why there's going to be a massive shift.

And I promise everyone out there, you won't have to talk to a human agent on a voice call soon. We'll get rid of it. Can you bring an iguana on an airplane? Well, it turns out for one airline, if it's 45 pounds and below, yes. For another airline, you cannot bring an iguana on. So they have different policies about iguanas, funny enough. So do you see that even for harder use cases? I mean, I'm thinking back to a chat I had with United.

I forgot my password, and I forgot some of the answers to those security questions. You're saying that you could build a product that involves no interaction between me and a person for even that kind of difficult situation. Yeah, and they're one of our customers. So we do a bunch of different use cases there on messaging and automation. But yeah, those are done. They may even be doing those. I don't know.

But I know they do booking reservations, changing seats. I think password resets are in there. But yeah, all of that can be done. We don't need to talk to humans anymore to get that done. Well, I hope we talk to humans sometimes. One thing on your generative AI products that was mentioned in your product launch is that it can do real-time fact-checking, which is something that ChatGPT has a lot of problems with. How is this possible and, more importantly, provable for your product? Yeah, so because we're in the enterprise, we call it enterprise AI, EAI.

This is a very specific type of AI. This AI, you know, has to be at the highest form of trust and the data sets and our platform and the models all have to work together to give a great outcome. So we have things in our platform that'll check, for instance, you may have conflicting answers. Well, that needs to get flagged and we cannot put a human in the loop then to say that's the right answer. So the beauty of our platform, though, is that it's humans that generated the conversations

originally. and usually those humans, the agents at the contact center, they're trained and they respond appropriately. So the quality of our data set, we have over a billion conversations we generate a year. That quality data set is what ultimately helps us drive AI into the enterprise. Get them back into the model and fact check. And the models say, that's not true. You say, is this true? And the models say, no. So you use prompts to then do the fact checking on it. So we have to do a lot of tooling in order to make this stuff work for the enterprise. It's not a straight API call. Part of this is a part of what I would describe and you've described is live nations,

sorry, not live nations, live persons. I hope that was a bad one. Part of this is a part of live persons, what you would describe as a turn, an inflection point. You mentioned this in your previous call and as shareholders may note that the stock is down more than 90% from 2021 highs. Can you provide some color on what it means for you and your company to be at an inflection point? Yeah, I mean, I've been through this a lot. And the company's had different legs in its journey. We had the, when we launched chat, made it through the dot com. We launched messaging first time, then added AI into it.

So it's the same thing here. I did a big restructuring of the company to sort of get us aligned to the biggest opportunity that we've been waiting for for 28 years. the ability for a machine to power these conversations at scale. And so we did a lot of cleanup. We actually got rid of a bunch of revenue and businesses that we don't think are pertinent to what we need to do. And when this appeared, it was like, we got to be ready to go for it. And so I think we're in one of the best positions right now to win the enterprise AI space because we've got over 3,000 customers, 500 are enterprises. We've got this amazing data set. we've got the technology we've got models we already have trust to deliver this to our customers

so i feel really strong about our position and it's gonna there's like 10 of us in the world that are doing this they're really bringing this technology to life and making money off of it so that that's what's kind of unique so when you describe getting rid of some of those spin-off brands i believe one of them is kasamba which was a allowed people to connect with fortune tellers across the globe. You've described, said that now you're focusing on recurring revenues. Is there more to that than just getting rid of some of those consumer-facing spinoffs? Yeah, we had some, we had like a, we call it a gain share labor business where we provide the contact center labor. We had a business related to COVID and COVID testing and AI. So there were some businesses

during the COVID period that came in that lasted and came off. And so those are the ones we kind of moved out. We sold the consumer business called Kasama. It's about experts doing chat. There's psychics, but there's also like tech support and all that. But it was very much a chat business, not an AI business. So we kind of looked at the portfolio and we said, what can power AI? And that's what we doubled down on. And we want to basically also run our company profitably and generate cash off the core. And that's also something that we're achieving now in Q2. The core business is now generating positive EBITDA. And that's a very important thing for us. Positive EBITDA, but your company has still had some issues finding profitability

on an operating basis, the profit from your core business. For investors really looking at that operating profit, without dates, not projections, but what's the story where LivePerson becomes profitable on an operating basis? Yeah. Like I said, the core business is it's starting to generate cash. We have a AI healthcare business that's a 100% grower. It's a big grower. And we're investing in that is taking down some of our ability to generate cash. But by year end, all of that will be generating cash. So we feel good. Once again, we look at the core businesses generating cash. We have an investment in a high growth business, which is called Wild Health. And then those combined by year end will be generating cash.

I want to move on to some trends. I do have to address one elephant in the room, and I understand if you can't respond to it. An open letter recently came out from Peter Feld in Starboard Value. Looking at the 13Fs, I believe this hedge fund has about a 9% stake in your company. They actually don't anymore. They sold it off. They don't? They sold it? They have about 3% now. They've been selling over the last few months, so they do not have a 9%. Okay. I appreciate that color because sometimes things change between December 31 and May 10th. But the value firm does have some issues with the company. It wants to introduce a new board. It wants you, frankly, out of the company you founded. It said,

quote, while we expressed a strong belief that LivePerson needed a change of CEO, the board insisted that it was committed to Mr. LoCascio as if it felt that LivePerson was finally well-positioned for success after years of challenges. Moreover, the company's stockholder unfriendly staggered board structure prohibited us from pursuing a level of change that would have resulted in a change in a majority of the board." It points out poor financial performance, troubling Glassdoor reviews, a loss of business focus. I'm going to stop there and just clear the floor. If you have a response, I also understand that you may not be able to. Yeah. First of all, we actually settled with them last year, strangely enough, and then half the board is new board

members. So we actually made changes. We restructured the company and got it to a place that I think most shareholders are pretty happy with. Stock's up today. We talked about the final restructuring, getting to profitability, positive EBITDA. This is what we did for 20 years. We ran a company generating cash, and we needed to get there. So I think we've done a lot of work that our shareholders want, and we've made a fair amount of change, and now we're just focused on growth. I think that's what it's all about, being focused on growth. Growth in where? In terms of gaining profitability, revenue, operating profit? Well, we have a 75% gross margin and stuff. So we have a good strong business and we have a positive EBITDA business and we'll generate cash

as an overall business shortly. But it's about profitable growth. If I could go back and change something during COVID. COVID grew our business tremendously. Every contact center shut down. They needed automation. We were there with the product and the platform. We had envisioned that. But we basically invested like it would continue going because we're in the middle of it. It looked like that. And it didn't. It came down, as we all know. So we needed to recalibrate. And we recalibrated quite quickly. But for 20 years, we generated cash flow in the company and positive margins. And that's how we've been here for 20 something years. So that's really the focus of the business. I don't want to break that rule again. I won't. It just doesn't make for a strong

company. And we got here and we'll continue forward with generating positive cash flow and EBITDA. All right. I want to move on to some fun and weird stuff, if that's okay with you, Rob. One thing I've been looking into lately, I started reading this book by Michio Kaku. It's called quantum supremacy and it's the idea that computers now are based on just ones and zeros and there's a there's a direct limit for what we can do in terms of artificial intelligence and conversations um fact checking there's there's massive implications in terms of biology energy you name it um i'm going to give live person a working quantum computer oh yeah what would you do with it i and i think this is it's interesting because like if you think about like gpus and

And there's a limitation right now on what we can process with these models. So I'm kind of in the same place you are, where the energy and the bottleneck is the GPU and how it works. And so I do think that there's probably a different way to do it. It may be quantum computing. I know there's a lot of focus on it right now because it is a bottleneck for getting us to scale these systems. on a really good cost basis. So yeah, I guess if I had one, I would rework the model

to be able to use a quantum computer to generate the outcome and the conversation. So I don't know if anyone's trying to crack the code on that right now, or they're just programming for GPUs. No, IBM, Google, they're all, I think it's sort of, well, you have the software race with artificial intelligence, natural language processing. There's, I would say a quieter hardware race to get to that quantum supremacy. Yeah. I even, I'm in a venture capital fund and they picked some kids out of Harvard. There's like two guys who are thinking of something different than a GPU to power this stuff. So I agree. It's like human beings have a pretty good result in breaking out of the limitation. And I think this is the biggest

limitation we have right now. I mean, the other part is that many years ago when I started the company, there's like three forms of input output for a human. There's typing, which is the internet represents that text and typing. There's talking and speech, which we're seeing right now. And then there's thought. And thought is the fastest process to gaining access to information and processing. when you think and you have to speak that's pretty quick when you have to type and think that's slow so that's why we naturally like to talk to machines it's easier for us but the next place is obviously the idea that i could just i could have a thought and that thought can basically come back with a response and when you think about thought most thought is in a conversation like we think in

conversations right so it's important that this what we're doing now will be something for our future. I don't know if I'll ever see it in my life, although I know Elon Musk is working on it. I don't know if I'll ever see it in practice. The Neuralink, it's almost a co-pilot for thinking. And we have to have that. All this will be helping us be better at what we do, but I think that's where it's ultimately going to go. Your company focuses on chatting in customer service and professional ways. There's another company that's private, but it does, I would say, some similar technology in terms of voice AI and chat AI for personal relationships. It's called Replica. And basically, it started as

a mental health application for people who are lonely during COVID, and they would introduce a partner for you to talk to. And then slowly, a lot of the relationships would become more flirtatious, and then the user would be able to essentially buy an artificial intelligence partner, it became more explicit in terms of romantically. And then people were feeling huge bonds. This was their girlfriend, their boyfriend. And then there was a huge tech switch where Replica basically turned off the romantic functions. And people were deeply depressed by this. People were lashing out, understandably. They changed the personality of something that they loved. You're in this space. I wanted to get your take on the replica story and maybe just

the human relationships with AI and if there needs to be boundaries set. And I mean, I've been thinking, I mean, the reason I feel like I'm here on earth is to, and my purpose in life, at least on the product side, is to bring something like this to life. I think the machine has to be empathetic and loving and caring, and we have to trust it. So I do think this is important. And one of the goals in the company is like developer her. You remember the movie Her? Yeah.

Better than Her. So Joaquin Phoenix is in a movie and there's a chatbot voiced by Scarlett Johansson. And Joaquin ends up falling deeply in love with the chatbot. That's correct. And that, you know, well, in the end, it's kind of like she's also everyone else's kind of lover. Yeah. She finds out. So not so good.

But I think in the end, the consumers have to own their own AIs. We have to own an AI that's ours, that's not given to us from a company. It may be like it is given to us, but it's ours. The data is ours. We train it. We have control over it. We name it. And I think ultimately that's where we need to go in the world.

And that's kind of my vision. And so this AI that you would own, I want to build out the vision for that. That's someone that could be maybe you would say a romantic partner, a professional assistant, a friend. What is the purpose of having that sort of, I'll say for lack of a better term, spirit animal? Right, exactly. That's really what is important is that spirit animal. I call it like it would be your pet, like your pet dog, like something you love and cares for you. And then I think the brands have to have their own AIs.

And I think these AIs connect. And so that's where we look at is we call it the digital twin, is that ultimately the brands have a digital twin of you as a representation of how you engage with their brand. And when your digital twin shows up, it can connect with theirs and they will get things done for you. And that's where I think this all goes. Brand has their connective twin, you have yours, and they're an arbiter. They communicate with each other and they get things done for you.

I'm literally having trouble envisioning this. What does it mean to like, what is a task or something that a brand in my, my digital twin, which I'm still, I mean, I had a, you know, I just, with my healthcare company, I just, um, you know, had a filling replaced and they, they said it was a non, it was an unnecessary thing. And my filling fell out and it was, and it's 500 bucks and, and they don't want to pay it. And I haven't had the time to even bother with it.

And I'm really, you know, like I know what it's going to, the pain of calling of this company, this company. So my digital twin, I should be able to say like, here's the stuff that I'll give you to prove the point that this was necessary and go take care of it for me. And that would take place because a lot of the big companies like healthcare, banking, telco, we all have this in our life. And these have the large contact centers and a lot of complexity to how to manage with them. And so the ability to do that easily and have my digital twin communicate with the digital twin at their company and get that done for me, that's really the key. That's really the key I see in life.

So I think that sounds great. I think the difficult thing, though, is to make sure- That's what I'm trying to do, by the way. We're trying to bring that to life in our company. Trying to bring it to life. And I think the thing that is difficult is almost making sure the digital twin stays is a pet and not almost the person becomes subservient to the digital twin. And I don't know how you would build that boundary in place, but that seems, in addition to creating an immortal

digital version of yourself, that would seem to be too difficult obstacles. Well, I just synthesized my voice. And it took about 300 training phrases to do it. And then I'm loading in all my knowledge, like all the written word. And there's a lot out there that I've done in my life, like these interviews and I'm loading it in, loaded in my Wikipedia page. And it's on our platform because you can load in all your data on our platform, even Wikipedia stuff. And, and now you can talk to it and you can ask it questions. And so in reality, when I'm not here anymore and my kids want to talk to me, or maybe my kids, kids, kids, they can, you know, talk to their great great grandfather and something's interesting about that and that's what i that's

what i'm trying to achieve i think this creates not immortality but some form of you've left yourself behind in a in a digital form in a digital form that's real like like my mom's still alive she's 83 and i want to do this with her you know she'll not be here one day but it'd be great if my kids could say you know grandma can you read me a bedtime story and and that's possible So it's exciting. We're not far away from it. No, we're not. I don't know if I'm delighted or terrified, Rob.

Well, the terrified part is if somebody tries to replicate you and use it in nefarious ways. If I take the video off and it's not me talking, this could be pretty bad. So I think security-wise, so I think there's this whole other part that has to get solved. But we're close now to doing things like this. and what it means if a brand really knows my digital twin and maybe some marketing tactics and privacy restrictions. I think that's huge. That's correct. Yeah. I want to dive into, I was presented with three industries where generative AI is going to be a game changer maybe in the more near future. We already talked healthcare, retail, and travel.

Those are two where I see I am a little less clear on the implications of generative AI. you pick one and then and then we'll dive into it uh i mean retail is an interesting one uh we have a we showed a demo uh when we just kicked off our new generative ai products that you call in and you're talking to this ai and you're asking for questions like and our example is said i want to get my mom a gift and she's 80 years old or whatever and says oh i recommend these gifts like oh yeah she loves jewelry and then we went down and down into that tree and then said oh i think she likes pearls. And then we said, send us a picture. And it sends you a picture right to your phone. So you're looking at it on your phone, the picture of it. They're like, oh,

that looks cool. Can you give me something that's between this price range? And that was all done really easily. And it's all real. This is not fake stuff. This is what our platform can do. And then you can say, buy it. I'd like to buy. And so on the retail side, it's going to allow for, I think, a different level of discovery. Because right now it's like web. And webs are like searching in categories or stuff but you kind of make up the category it's like my you know you say my mom 80 years old lives in new york you know help me and it can it can go well based on that the other thing we did is i'll tell you we did another thing is that i we downloaded like uh my my whatsapp uh messages with my wife and we tested like um i want to give my wife a gift

i'm traveling and said oh your wife loves horses yes i've got some ideas great and then i was in London. I was like, well, where can I buy those? And so it's really interesting. If you can let the AI get in the middle of those conversations, it can think of things that you may not even think of. And I think that's pretty interesting. So in retail, it would maybe like it fits in as a personal shopper. As a personal shopper and discovery. And I think the discovery elements are so much more detailed. Like if you look at Google today and you put in, I go in there and put, i want to buy a gift for my 82 year old mother it's going to give me all this bs websites like gifts for mom you know yeah then you're going to get dropped on some other retail site but the true

it went it what's great about is it went from like flowers uh yoga mat uh jewelry it picked these things that are very disparate but for that age group it was like right on and then i was like okay cool and then i said jewelry and it's like um what does she like she likes pearls and we went down that road and then went down pricing. So it's really hard to do that with the AI previous degenerative. And it's hard to do that on search and stuff like that. Well, it's the reason it's tough on search because you're filtering through ads. Yeah. But to me, there's no reason that you couldn't introduce ads into the AI platform, right? If you're a brand and you have Rob's digital twin hey i'm out here selling uh i'm out here selling shoes uh let me let me present this

to to rob at the right place in the right time right and that's that's ultimately we know google is going to put you know like they announced today they want to put uh generative ai into their search microsoft's doing with bing um so yeah i think ultimately discovery um is going to be very different and then once it understands the digital twin and once again i put in information now in my conversations with my wife, I'm putting in these very personal conversations saying, figure it out. It figured it out. Now, I wouldn't give that to Google. I definitely wouldn't hand that to Google, that information. So how do I get this localized with my called digital twin? I want to talk about one other kind of weird angle with AI that has my attention.

I know that you, we talked about this a little bit in the pre-interview where I think you have an opinion on it. But one with natural language processing that I don't think is getting enough attention is interspecies communication. I think this is going to be the equivalent of the atom bomb or the internet if slash when it happens. And a lot of what's being discovered was already known by indigenous peoples. And this is a lot of renewed discovery. But now because of these language processing programs, we know how animals talk in much greater detail. We know that bats speak in motherese to their young. They change their pitch when they're talking to a little bat, and that bats distinguish between genders. This was known a while ago, but the bees can wiggle and show other bees exactly where to pollinate. And then you can have robotic bees come into a hive and then show them pollination sites, but it can get much darker.

You can imagine a fishing boat saying, hey, tuna, hey, school of tuna, get over here. It's used for precision hunting. Your thoughts on this unlock, am I overstating it? It really isn't something I've been thinking about. I know you brought it up at the beginning, but I do think it's quite clever. The translation between beings and how we translate obviously will have a key part because there's a lot of good and bad that can happen with that. So I thought when you brought it up, it's quite fascinating if we could truly understand.

And I guess in some cases we do the language of another, you know, species, species. So I don't know what will happen with it. It'll always start with pets. The pet market is so frothy. so like it's going to start with can we communicate with our pets better and understand let them understand us better it's pretty cool if you could really get your dog to understand like what you're feeling you know and and and uh what you needed to do but uh yeah it'll be interesting i think the pet market it's probably a pet market thing because that has billions of dollars into it

that's how usually commercially things drive things you know i was talking to a long time fool mac greer about this and the thing is is i don't think you want to know what your pets think because it's probably not as cheery as what you'd like it to believe. No, and you know the old joke, like there's aliens looking on Earth and they see humans basically walking dogs and picking up their poop. And they're like, wow, look at these species. They have these people who pick up their poop. They must be really advanced.

So it's like maybe the dogs from on the show, you know? as always people on the program may have interests in the stocks they talk about and the motley fool may have formal recommendations for or against so don't buy stocks based solely on what you hear i'm mary long thanks for listening we'll see you tomorrow Thank you.