Chatbot and the Future of Conversational Intelligence: Sonny Patel

Auto-generated transcript. Speaker names, spelling, and punctuation may be slightly off.

Mark Evans: In the technology world, hyperbole, also known as hype, is inescapable. Anything new is touted as the next best thing since, well, the last next best thing. There's overwhelming enthusiasm and expectations are sky high. And over the last few years, artificial intelligence has received this treatment. You'd swear that AI was a gift from the business gods given the possibilities that everyone has touted. But like everything, AI has thrived in some areas, and it's still a work in progress in others. To delve into the world of AI, I'm talking with Sunny Patel, SVP conversational intelligence with LivePerson, which develops conversational commerce and AI software. Welcome to Marketing Spark.

Sonny Patel: Hi, Mark. Thanks for having me.

Mark Evans: My description of LivePerson, I cut and paste off the web, so it's pretty formal. Can you talk about what LivePerson does and as important, provide some context about the role of AI within the software?

Sonny Patel: And I can also start from where we began and how we evolved. So LivePerson at this point is a twenty six, twenty seven year old company. We were our founder and CEO, was a pioneer for the web chat idea. So back when the.com back in the .com era when everybody was building websites, he essentially pioneered this idea of why do I need to navigate through this very artificial, you know, looking website. I should be able to just talk to a person like we were used to until at that that point. So that's what we started out as and we evolved. Now more recently, we are investing a lot and building a brand around being a conversational AI company. So a quick fun fact is that Fast Company has named LivePerson as the number one most innovative AI company in the world.

Mark Evans: So when you say conversational AI, I think a lot of people equate that to a chatbot. Is that basically what we're looking at when it comes to what LivePerson does?

Sonny Patel: A lot of the companies that are in the conversational AI space, that's what they're focusing on is just being able to build chatbots that probably tackle some of the more simple use cases like answering frequently asked questions. We are innovating beyond that in the sense that not only do we want to enable brands to automate the conversations that they have with their consumers when the consumers are reaching out for any reason to buy a product, to inquire about a product for a service related question, we also want them to feel that human touch. Make that conversation feel more empathetic, feel more human like. And the way we do that is by understanding first what is the consumer really intend to do during this conversation. And with and and in order to do that in a very effective manner, we need to understand, we need to measure, we need to remember, and then use all of that knowledge to personalize the response that we are providing back to the consumer. So to me, that's what conversational AI does and that's the power behind it, not simply, you know, responding to a question saying, what are the hours you're open and answering back with a static answer no matter who's asking the question.

Mark Evans: It begs the question of how big of a challenge, how much AI power is needed to answer questions quickly and effectively. And as you say, you do it in a very personalized, empathetic human way because robots tend to talk like robots and humans talk like humans. But if you can get a robot to talk like a human, obviously, you're gonna have a better experience. So you've got a big team. I think you head up a 200 person team. So live person must be investing a lot of money and a lot of a lot of intellectual firepower to create that that human like chatbot for lack of a better term.

Sonny Patel: Very much so. And it all starts with data. Right? And I think that this is definitely definitely the distinct advantage LivePerson has. We support over a billion interactions on our platform every month. And all of that data essentially gives us the foundation to understand how these conversations are going, what's going well, what's not going well, where can brands do better. And we use all of that data to model and feed into our AI powered capabilities. So so very important, and I think that that's no one can build an AI or get better at it without that underlying data, which given our history, given the customer footprint that we have, we have that distinct advantage. And in terms of the investments, where we are definitely more focused on is developing better understanding of where the conversation is going well. Obviously, when something's going well, we do more of it. And where the reception from the consumer is better, we give those signals to the brands. So brands also understand what's going well, do more of it, and also give them signals on areas where they should be putting in more effort, either by improving the dialogue or when, it's an appropriate time to pull in a human being into the conversation and inform them. And the idea is to also automatically take action and not just wait, not just say, hey. Here's a measurement of conversation not going well, but make recommendations on what they could do at that point.

Mark Evans: In the marketing world, there's a lot of emphasis on personalization these days. I guess one of the problems is that marketing automation has allowed marketers to do things at scale, but it's come at the cost of everything seems very generic these days. Other than if insert first name here when you're doing email marketing, for example. The experience seems very impersonal, and a lot of consumers are now sort of becoming very aware of that, that they're just being part of the machine. Can you talk about the value that LivePerson brings to brands and marketers, allowing them to do personalization at scale in a way that is empathetic and appears to be the right message for you at the right time.

Sonny Patel: The word that's evolving around that concept is hyper personalization. Right? And this is a good analogy is, you know, how we interact with other human beings. So, yeah, I mean, you and I are meeting for the first time. So, obviously, our conversation today is gonna be very generic. You know, it's gonna be a little different the second time we talk to each other. It's gonna be very different the tenth time we talk to each other. And it all depends on the fact that it's a two way conversation. You know, when you have questions about me, you ask me, I ask you. Over time, we develop an understanding of who it is that we are speaking to and use that knowledge to, you know, respond appropriately based on the preferences that we have gathered. And that feels real. Right? So the the injecting a name, but keeping the rest of the message the same, even though we call it personalization, but it's not really. Right? So and I think that this is the distinct advantage conversational AI has over traditional AI is that in conversational AI, for most part, you know, we are going through the conversation, gleaning preferences for sure, and, you know, combining a combining a lot of that with the signals. But we could also just directly ask the consumer when we are talking to them and basically say, hey. Would you prefer x versus y? And use that explicit knowledge to improve the future interactions. And that is something that that is something where that's is that's an advantage of these having conversations as a basis for developing the understanding of who it is that we are speaking to. The dialogue that develops as a result of it looks very different from the more templatized personalizations. The other aspect of personalizations which people are starting to get very wary about is this whole notion of companies using cookies. I think people in general feel very achy about being followed, you know, as they're surfing through the Internet. And that's again the consent based personalization is also something that we can do on during the conversation. And the third thing is about the when we use personalization. Right? Personalization for the sake of using it can sometimes actually backfire. So I'll give you an example. I pick up my kid most of the most of the days around three three 3PM, I leave from home. And over time, my maps application learned that that's where I'm going to. And, you know, makes a suggestion saying, hey, you know, I got into the car, it knows I'm in driving mode, and it automatically plugs in where it is that I'm driving to. Very delightful because it's actually helping me, you know, bypass a few clicks that I would be doing at that time. When AI is used to personalize an experience, it's it works, it's fantastic. The fact that my phone, that application tracked, that that's what that's the pattern I was following doesn't really bother me because I'm getting a benefit out of it. I think where the variance for the personalization for being followed comes in is when the AI quite doesn't work. And I think that within conversational AI, we have the opportunity to inject personalization at the most appropriate time. And, also, again, we can get a signal directly from the person that we are engaging with whether they liked it or not because we are actually hearing back from them. And we understand whether the conversation from that point onwards had a more positive and or a more negative sentiment attached to it.

Mark Evans: I'm gonna take a step back when you talk about the idea of hyper personalization. In simple terms, is that the killer app for conversational AI? The idea that or is it part of the killer app recipe? The fact that the conversation that I'm having with a chatbot or an AI powered software is very is all about me and my experience and my the things that I wanna do. Is that is that the killer app when we're talking about conversational AI?

Sonny Patel: That that would be the vision of it. Is that let's say you and I reach out to the exact same brand, and we start to look for products, something to buy. The experience that is presented to you should be different than assuming with different people with different preferences should be different. Also in terms of even little things like when is a recommendation presented to you? At what point during the conversation? The fact that the AI or the chatbot is able to detect and do it accordingly to maximize, one, your satisfaction because you're the consumer here ultimately where you feel like, wow, that was the right time. It did ask me a question that made me feel valued, made me gave me value out of that conversation. And also the impact of that for the brand is that you're more likely to shop with them again because you feel like they know me and they're personalizing it to me as opposed to having a script. And one of the statistics that we have discovered is that 79% of the consumers prefer interacting with brands where they feel a personal connection. So the the value of why are we investing so much in enabling brands to present that personalization is that it overall increases a lifetime value from all of the consumers and increases the likelihood that those consumers will actually go back to the brand, engage with them again.

Mark Evans: From a high level perspective, when we look at the impact of AI on marketers and marketing, where are we at in terms of the journey? Like, how much value are marketers getting from leveraging AI in terms of the software that they use? AI for the sake of AI, I don't know. If it doesn't make your job easier or allow you to do things faster or better, then there's not much value in it. So how much of an impact is AI making on marketing right now? And and where do you think the the areas that it's making the most impact are?

Sonny Patel: Yeah. And I'll I'll I'll focus a bit more on the conversational AI aspect of it because I I think it's incredibly powerful. This this notion of engaging, but engaging with you to have a two two way conversation as opposed to just blasting marketing messages, which we know traditionally are not as impactful. Right? But, basically reaching out and saying, Mark, I think you would like this product or or this is a time to kind of renew, your subscription or, you know, here's your opportunity to pay a bill, and then you can take action based on that that initial message. So what we are finding is that about 83% of the consumers want to interact with brands to to buy products via messaging. I'm one of them, by the way. I hate making phone calls. I'm not as much of a in person shopper. I do I would rather do it online. I would rather do it over messaging when I have a question as opposed to picking up the phone. I hate phone calls. Also, out of those 83 that are willing to engage, 7% overall of consumers are willing to make a purchase when the brand supports messaging. So there is a lot of power in having these interactions over messaging, which is at the convenience of the consumers. And I think that the AI aspect of it for marketing is, again, like I said, rather than have this genetic blast, like a marketer wakes up in the morning and says, today, I'm going to talk blast this deal for toaster, which how many of us are looking for a toaster on any given day? As opposed to, you know, again, personalization, using machine learn models to vary the recommendation based for each individual based on their past purchases. That is very impactful. And then do that over do that over a conversation, do that over a messaging. Each of that is very, very incredibly effective.

Mark Evans: Obviously, there's been a lot of hype surrounding AI and its impact on how companies do business and interact with customers. So the big question, and it's it's a loaded question, is has AI lived up to the hype, or have expectations been too high, or are we still early in the AI game, keeping in mind that this is not new. AI has has been been working on for decades. Provide some context in terms of the fact that a lot of people think that AI is a disruptive game changer that's gonna basically upset the apple cart aligned with the fact that maybe AI is it's seeping its way into the business world, but it's still early days.

Sonny Patel: I think we're very early. We're still experimenting. There's obviously, like I've mentioned, there's a few applications of AI that have proven to be incredibly, effective. But I feel also feel that we have had a few false starts. Right? And by the way, before joining LivePerson, I was, at Amazon working on Alexa, the the voice recognition piece of Alexa that was, something I helped launch. And, you know, we had this notion of these AI assistants being omnipresent, being able to understand our every need and making it so easy to have a conversation and, you know, just make our life easier. And I don't think that has quite panned out. With all the Alexa's I have, in my house, most I use them for is play music, play sleep sounds when the kids are sleeping, or turn lights on and off, which is incredibly convenient. But for most part, it's been me saying, Alexa, do this. Alexa, do that. But it's not been as engaging. I didn't understand the question I heard. That's my Alexa waking up. So

Mark Evans: Okay.

Sonny Patel: Sorry about that.

Mark Evans: See, it's real time AI right there.

Sonny Patel: So I think that this vision we all had of the world we would live in, like this black mirror world where we are gesturing and interacting with all the devices that are around us all the time, that has not quite panned out. And I feel that this is where obviously, we need something to interact with, so having those devices is great. But make that a conversation as opposed to command based interactions, I think that's the that seems to be the gap. And that's for that to actually work, it can't be a small subset of devices. It has to be more it we need to be inclusive of more devices that we that are around us. And also, cannot be purely command based, but it's more integrated as if we're interacting with the human beings around us. And this is something that LivePerson is definitely thinking very deeply around. We launched quite a few innovations under the branding of Curiously Human. And the whole notion is that when we have AI we have these interactions with AI, whether I initiate the conversation or the AI initiate the conversation, it feels as if we're interacting with human beings as opposed to distinctly understanding that, hey, there is this thing called AI named x y z that I'm interacting with.

Mark Evans: If we flip the conversation back to the chatbot landscape, From the outside looking in, it appears crowded and noisy. There are dozens, if not hundreds of companies playing in the chatbot world, chatbot sandbox. LivePerson is obviously a fast growing company, so it would be interesting to get to get some perspective on the market's growth, its evolution, some of the more exciting trends that we're seeing within that world. I mean, obviously, from my perspective, from a lot of consumers perspective, we see that little box in the corner. When we have answers or we get frustrated, we hit the box. But step back and provide some big picture perspective on how you see the market moving forward. And are we gonna live in a world where there's hundreds of chatbot companies battling for companies' dollars and cents?

Sonny Patel: I think that's where we are is that there's a lot of companies that are tagged as conversational AI companies, but they're really chatbot companies. They're really providing tools on how to build chatbots and basic chatbots at that in terms of, okay, so load up your what are the questions that are most frequently asked? Load that up. Set up a chatbot to do that. And where LivePerson, definitely, we distinguish ourselves. And I think that that's a that's tackling the true power of AI is to think about the entire journey. So sure, there is an aspect where you build a chatbot. Sure. You can start with a more basic chatbot, but there has to be an evolution to make that automation more sophisticated. And in order to do that, you know, you need to bring in some conversations. So start to build some data. Use the that data to build an understanding of what's going well, what's not going well. So it, like, defines some metrics even to understand how do we measure a good conversation and then build an so measure that. And using those metrics, build understanding, develop some insights to say, oh, when the customers are calling about basic questions, the chatbot is doing great. It's all fine. But when the, you know, when the consumer has a more complicated question, like, that they want to provide some feedback on a product that they bought. That's where the chatbot is failing. So those are the these are the intents where the chatbot is not doing well. You know, provide that insight back to the brand and also go with some recommendations to say, okay. When these intents are being brought up by the consumer, here are the more nuanced responses that you can get back with them. Here are the here's where personalization can be provided. Or in a in a very complicated case, that's when to bring in a human being and provide recommendations for actions. And in some cases, if you are able to take an action on behalf of the brand, just do it. And to me, that is what a conversational AI does. And when we talk about a lot of companies and the space being crowded, I feel that they're basically starting this the first step, which is essentially helping brand set up an automation and serve some basic use cases, and we need to go much beyond that.

Mark Evans: You raise a good point in terms of measurement. And I did wanna ask you about that to get some insight into how companies quantify the performance of a chatbot, the impact, the benefits, and as important, the ROI. I mean, a chatbot is a great tool to engage with customers and provide a different or better experience. But at the end of the day, it would be interesting to know from your perspective, from a live person perspective is, you know, what does that chatbot dashboard look like? When I look at my KPIs in terms of how the chat part is performing, what are the different metrics, and and how do I get actionable insights so that I can look at the metrics and do things better?

Sonny Patel: Yeah. And that's the conversational intelligence space. That's what we are looking at. So so here's what we do at LivePerson. So we take a number of different signals, look at a number of different metrics, and we use that data to really build an understanding of four distinct measurable areas. On the consumer side, we want to measure the effort. How much effort did the consumer put in in order to get their intention fulfilled? Whatever that that it was that they reached out to the brand for. And also, at the end of the conversation, what was the emotion? Like, what was was it a net positive conversation, net negative conversation? So that's from the consumer side. Again, we distill it down to these two metrics. However, there's a lot of data points that go into measuring those. And from the brand side, we want the brands to develop an understanding of how effective were they how how efficiently did they respond back to the consumer's intent and also how well they handled it. So it's efficiency and effectiveness. So we call that the four e framework. When we think about what is that feedback we are giving to the brands that use our platform, we want to distill all the distinct data points into these four areas of measurement that they can take a look at and understand where they need to take action. And the areas where we are looking to innovate and build more capabilities is that once we have those measurements, what actions should the brand take? And instead instead of putting the onus on the brands themselves or, you know, the the agents or the managers who are managing the contact centers, we wanna be the ones to make the recommendations just to make their life easier. So the AI we are building is not only useful to provide personalized responses to the consumers, it's also something that we want to use to the advantage of the businesses so they can operate more effectively.

Mark Evans: One final question, where can people learn more about you and LivePerson?

Sonny Patel: So we have our website liveperson.com. A lot of what we do is on our website. I would also recommend following us on LinkedIn. We are constantly updating our audience on the new and improved enhancements that we make, and I'm on LinkedIn as well.

Mark Evans: Thanks for your insight into the exciting world of AI and chatbots. It feels like it's early days. It feels like we're gathering momentum, and the chatbot world will look a lot different a year or two years from now. And I wanna thank everybody for listening to another episode of Marketing Spark. If you enjoyed the conversation, leave a positive review. Subscribe via Apple Podcasts, Spotify, or your favorite podcast app, and share via social media. And to learn more about how I help b to b SaaS companies as a fractional CMO, strategic adviser, and positioning and messaging development, email me at mark evans dot c a or connect with me on LinkedIn. I'll talk to you soon.