Hello, everyone. Welcome, and thank you so much for joining us here today for this webinar brought to you by CX Network and our partners over at Coveo. Today, we will be looking at how you can use generative AI to really elevate your customer's experience, and improve self-service as well. So with me to go through that with you is Devon Paul, who is senior marketing manager for services, and Patrick Martin, who is general manager of service solutions both at Coveo. Before I hand over to them, I just wanna make everyone aware of a few housekeeping points. So first of all, please do be aware that we do have a q and a box, which you can use to submit all of your questions for Devon and Patrick as the session goes on, and we'll get to as many of those as we can at the end of the session. And secondly, I'd like to also make you aware that there is a case study in the resources box, that you guys can download for free. It's really interesting, so please do go ahead and check that out, after the session has finished. You can also access that via the lobby. And finally, this session is being recorded and will be available on demand. So those of you that are joining us live today, if there are any talking points that you'd like to revisit or if you'd like to share this with your colleagues, then you will be able to do that via the event landing page on the CX network site or via the Zoom lobby. But for now, it's my pleasure to hand over to Devon and Patrick. Thanks so much, Chloe, and welcome, everyone. So excited for you to join us here today. Patrick, my partner in crime, always good to have you and speak with you as well. Same here. It's the it's Devin and Patrick show. Yeah. Exactly. Well, we we do have a great show lined up for everyone. And as Chloe said, right, we want it to be as much of a discussion as possible. So please, by all means, hop into the q and a box, ask questions. We'll answer some of them live. We can answer some, via text. We won't share anyone's names or details, so don't worry about that. But, you know, some of them you wanna take offline and have a follow-up conversation. Right? Patrick and I love to talk as you'll, quickly find out here today. But, we are talking about a topic that, you know, some of you might be sick of hearing about by now, but then again, you willingly and freely joined this webinar. So, you're at least somewhat interested. It's generative AI. And specifically, how are we gonna leverage generative AI to drive key business results for self-service? Now I I've been working in the customer support world for a little over fifteen years now. And every single year that I've been studying this space, getting our customers to self serve ranks in the top three of service leader priorities in some way, shape, or form. And in twenty twenty four, it's no exception. Right? It it is, how do I get people to self serve? Patrick, I know that you've been looking at this for a long time. In fact, you used to, run, the the service function here at Coveo. And what have you seen in this space? Well, excuse me. The, you know, cell service has been around, like, ever since the the Internet pretty much. And, you know, there has been a lot of a lot of focus on cell service. And as the industry transitioned from on prem to cloud, the approach to self-service completely needs to to transform. You know, prior to that, the on prem business is the the customer owned the asset, and a lot of companies, especially in the tech world, leverage self-service as a way to, I'll say, put a barrier between them and their support organization. They they hid the email. They hid the phone number. They really didn't want them to talk to to to their support agents, and they threw a lot of information out there, a documentation site, an online help, a support portal, and then chatbots came along and all of that. And as we carried over to the cloud, now we're service providers, and the customer has the luxury of choice around the channels that they wanna use, whether they wanna go self-service or not. And as we'll see later in the webinar, there's clearly an appetite for self-service, but it it can't be a disjointed experience. And that's probably one of the biggest struggles that we see with with our customers and and in the industry is how do we go through this? You know, we talk about digital transformation for a long time, but how do we go from this, you know, siloed dysfunctional self-service ex experience to really leveraging technology like generative answering or GenAI to, kind of have a coherent, unified experience and think about it from a service journey, not just assisted and self-service. So, yeah, we'll touch on these themes today, but that's that's one of the main things that, you know, we we see with our with our customers, and that's been talking around in the industry. Yeah. That that that's it. Right? I've long said, customers are like water. They will find the easiest path downhill. As long as, there's a channel for them to use, they're gonna use it. The the first ever research study I worked on way back in two thousand nine, posed the question, how in the world can we get customers to try self-service? And then we went out and asked tens of thousands of customers. Turns out you didn't have to try. People wanted to use it. It was getting them to stick. That was the problem. How do we get them to resolve their issues? Now, two main things have changed across the last decade and a half. The first is, of course, technology. Right? Technology is ubiquitous in our lives these days. We're all of it all of us using it day in and day out for most of the the task, and the technology has gotten much better. Even more so, with with the onslaught of generative AI that that's come in the last year and a half here. But what's also changed, you know, is not just technology, but psychology. Human psychology has changed in a way that people seek fast, easy answers. And across the last four years, in particular, we've all become much more digitally. Right? We got forced into doing things on our own. And for the vast majority of customers, they said, I like it. I want more of that. Please, you know, help me to do that. But what is always true and will always remain true about customer, psychology is that people just want the same thing. Right? Customers only want one and only one thing, which is resolution. Just fix my problem. Right? Make the pain go away, and I'll be happy. So with that in mind, let me start by setting some context here around, you know, the the human behavior and psychology that has changed and exactly, you know, why we're being more driven toward self-service today. So here, we're just looking at three stats on the page. Could have easily pulled hundreds of stats and overwhelmed you. Right? But what what they're showing is that customers are far more likely to, you know, try and and, prefer self-service today. So moving from left to right on the page, you can see from a study that Gartner did, seventy percent of customers say that they prefer self-service over contacting a CSR today. Doesn't mean they'll only ever use self-service, although point coming on that in a second here. But the idea is that customers are mostly digital first. I'm going to start there. I'm going to attempt to self serve. And even if I can't, I wanna gather as much information as I possibly can. So that's a lot of where that preference is coming from. In the middle of the page, right, we know that, executives are feeling this as well. The folks at CMP Research, did a study that found, almost three quarters of service executives said that there's been an increased demand. Right? So customers want it. We're feeling that pinch. And if we take a peek forward at what future customers are gonna look like over on the right hand side, we can see that that this pressure is only going to increase. This is both b to c and b to b metrics. It was a a sort of pan industry study that were done here, but we're feeling this pressure across both to the the comment that came into the chat. Thanks for that. But if you if you look over on the right hand side, almost forty percent of Gen zed customers, said that they will give up on a service issue entirely if there's no self-service option for it. It's like, I'll just let this pain nag at me until I can leave you. And if you're thinking, this is just the kids, they'll grow up. Well, thirty eight percent of millennials and somewhere in the high twenties, of Gen Xers said the same thing. Right. So this is a pervasive problem that's happening. Customers are demanding more self-service, and service leaders are looking to respond and respond using technology solutions. Let me show you what I mean. So we're responding with everybody's new favorite tech capability, of course, generative AI. Right? So looking across the the stats on the page here, you know, almost ubiquitous, positive answers to a BCG survey of service leaders where ninety five percent said they expect customers to be served by an AI bot at some point in the service journey by twenty twenty five. Right? And the keyword there is service journey. It's not one and done for our customers any longer. It's a journey that we're looking at. In the middle, the folks at CMP research, again, eighty two percent of companies that they surveyed expect to offer generative AI in some way, shape, or form in customer facing self-service by twenty twenty five. And eighty six percent of companies expect to be using this in their knowledge bases by twenty twenty five as well. Right? And so as we look at this, you know, Patrick, I know what we've talked about this, you know, where we we've been looking at Genet for a while, but you you've come up with a really, cool way to to think about twenty twenty four and and what we're calling it. Yeah. Well, you know, we can clearly say that twenty twenty three was the year of of hype. Right? When CHAN GPT came out, you know, it it it got everybody's attention. It really took the world by storm. And then everybody kinda said, well, we want this. We want that. We we see, you know, the benefits and and all that. And there was a lot of lot of noise, right, that came out through the year. Everybody has a GenAI offering. Everybody is gonna do this. Everybody's gonna do that. But in twenty twenty four, what we're gonna see is is really the year of of operationalization of this technology. More and more vendors are coming to market with actual working solutions in various use cases, and support is clearly an area that we see the gains both from a productivity standpoint. So anything that can enhance the the agent experience and being more efficient, leveraging GenAI and things like that to to do, you know, mundane, repetitive tasks around case summarization, article creation, things like that, but also being able to quickly, you know, document or find answers without having to read through documents, having to search left, right, and center. So, definitely, that's one aspect of of, you know, where we're seeing a lot of focus. But the self-service aspect is also important because where we're going with this is the ability for customers to not just search for information and try to resolve their problems, but actually get the answers that that they expect, which will have an impact on what is coming in our support, support organizations. Right? With with this technology, we're expecting to see a significant lift in terms of self-service success, which means whatever's coming into the to the support organization will be more complex, will be unknown, which would have a beneficial impact on employee the the overall employee experience because we know that one of the main factors of agent attrition is not being challenged enough and having to deal with the same issues over and over again. So we clearly see the coming of Gen AI across the entire self-service and assisted journeys to be a positive for both customers and agents as well. Yeah. It's so true. Agents wanna be challenged. They don't want you to make their jobs harder necessarily, but challenge keep me engaged with things, especially if, your CSR population by the way, I'll use the terms agents and CSRs interchangeably for folks on the phone. They're they're the same humans I'm talking about. The ones answering calls or chats or emails, on behalf of your company, that they wanna be challenged, because we're we're hiring smart people in often cases, especially these days. There's a great comment in the chat here of, you you know, let's not forget that of the people that say they prefer self-service, that they don't wanna feel devalued. Exactly our point. We're on the the same page, which is it's not a either or situation. In fact, when I used to, wear an analyst hat, I I would get questions all the time of, so we wanna turn a bunch of our customers, you know, the millennials, they only ever wanna self serve. So we want them to be self-service only customers, and I'd say, no. Stop right there. That's gonna kill you. Right? Because what happens when you think about it that way as self-service only customers is there are situations that these people are gonna run into that can't be self served. Not every issue is good for self-service. But increasingly, and with the power of generative answering and Gen AI, more of the issues that we are solving can become self-serviceable. And that's good news. That's good news for us as a company because it helps our bottom line. It's good news for our customers because they get faster, easier access to things. And if it's done right, which is what we'll show you going forward here, Patrick used the term unified earlier. Man, that's gonna be important in the world going forward. Because if we do it right, it's beneficial for our people as well. They get augmented. They get a powerful tool that helps them to be more successful in their job. It helps to elevate or or professionalize the the customer service function in many ways. Right? It's let's not forget that this is a function that started in the, I forget which CEO had said, you know, the necessary evil. We don't wanna talk to our customers, but they keep calling. So we have to staff some people to pick up the phone because there's problems with them. Right? So that's what we're we're avoiding here. Yeah. And if I if I if I can piggyback on what you one of one of the things you said, Devin, is how we've been thinking about, you know, the customer experience historically is we've been thinking about it from a digital standpoint or from a self-service standpoint saying, oh, there's this self-service journey or self-service experience here that might be owned by a digital experience team. And then there's the assisted experience, which is, you know, the call center or the support center or whatnot. And we we kind of created a kind of a gap in the experience by by thinking about it separately. I think with Gen AI and the coming of Gen AI and, you know, what you mentioned is that the phone will always ring. There will always be situations where someone wants to be served by a human being because of an emotional situation or whatnot. Think about it. You look at your credit card receipt, and you just notice that, you know, you've got ten transactions that you didn't do. You probably don't wanna self serve in that in that situation. You probably wanna talk to someone who's gonna reassure you and make sure that, you know, you're they're gonna take all the right steps to, you know, erasing these transactions, canceling your credit card, sending you a new one. You know? Yes. You might you might some people might be might feel comfortable self serving, but I would think that you wanna you wanna serve that person, with a more, you know, personal touch. So I think it's it's time for us as an industry to think about the service experience. Yes. And that encompasses cell service and assisted. And think about it. Like, the customer now has the luxury of choice in terms of which channel they want to interact with us with. They will have a preference for self-service. The data clearly shows it, but you can't forget about the assistant channel and how you connect the two together. That that's it. So long. It it's lived in these two separate worlds, and that just can't continue for us. Because to your customers, you're one company. That's it. One company at the end of the day. Now, of course, if it were as easy as push a button, install generative AI, Patrick, we we'd be done nineteen minutes in here. But Are you saying it's not that easy? Yeah. It's a muddy, muddy landscape. In fact, it's just looking at, this is one graphic of looking at some of the types of, generative AI that are in the customer service space. Right? There's content creation. There's content discovery. There's conversational. There are AI simulations. Right. So there there's lots of types and flavors of generative AI at our, disposal here. But let's not forget about this. Right? At the end of the day, your customers want the same thing they always have. So back to where we started. I want an answer to my problem. Right? Give me some sort of solution. Information that's gonna get me back to using your product, or get me on to the rest of my day, other things that I have to do in my life. Right. So that's why we focus so much here, at Coveo, on this idea of generative answering. Right? When it comes time to deploying GenAI capabilities, plural, right, we wanna focus on generative answering. So, let's take a look at what that is, at least in our definition here at Coveo. Now I'm gonna assume that everyone, on the line here can read. So I'm not gonna read out the the definition here, but you can see what we're calling out as important, around the outside here. Right? It's gonna be conversational. Right? Customers may have multi part questions, and they they may need to have some sort of conversation or ask more than one things where the context of that thread flows with them before they achieve resolution. Or there's a next question that they don't know yet to ask. And that's, something we wanna pick up on here. It's gotta be, in our opinion here, unified. Alright? Enterprises tend to have content that lives all over the place. And if that content, isn't unified, if there's not a single source of truth, so to speak, then what's gonna happen is that customers are going to get inconsistent information. Right? And nothing drives unnecessary calls more than I got two different answers from two different parts of your website or my digital experience. I'm gonna throw my hands up, and I'm gonna call a human. But I'm probably gonna chat you at the same time as well. So now I'm tying up two different channels at the same time because none of us have ever done that before. Right? If you move to the bottom and the bottom left here, right, it's about utilizing a combination of things here. Semantic search, relevance AI, prompt engineering. Right? Combining those things together to, you know, sing in harmony with with one another. Finding the right pieces of information based on who that customer is, right, within the context of their journey. And then bottom middle, it's gotta be grounded. Right. One of the biggest risks of generative AI is that it can hallucinate. Well, if you are grounding that answer in your company's fact base, then the the risk of hallucination goes down significantly. It's never gonna get to zero, which, by the way, is the the same with our humans. Right? People on the phone, they may not be hallucinating. At least, I hope your CSRs aren't hallucinating at work, but that's a bigger HR issue. But the idea is, like, they can give misinformation. They can give a wrong answer at times. Right? And we're never gonna get that risk down to zero. But when you ground it in your fact base, it's going to go significantly down. And then on the bottom right, it's about personalization. Right? Customers wanna get answers that reflect their own situation. And the, example I always run to here is imagine you're an airline, and you've got two different customers asking, what's my baggage allowance? Well, one of them may be one of your top tier status members, and they've got a different baggage allowance than someone who flies once a year and is just on this airline with you this one time. Right? And if you give either of them a generic answer or it's a long document that I've gotta scroll through to see where I fit, it's too much effort as it is. Or if you give both of them the same answer, one of them is gonna be wrong. Right? So that's what we mean by personalization. It's finding who the asker is, what the intent of that user is, and pulling that into play here. Alright. So, that's the the way that we want to start to think about this here. And so thinking about it, right, it's not just one flavor of AI that you're gonna use here as well. Right? It's gonna be many different types of AI that you're gonna bring to bear to impact your customer journey. Right? So to some of the the chat that's coming in here, which I love, keep that up, by the way. You know, the the idea, that we we need to create this unified thing, this unified experience for customers all the way from, you know, websites or even, Patrick, what what I'm missing here is in product. Right? If you're a SaaS provider and you wanna have help and support right in your, SaaS product, you can do that as well. But interjecting the the relevant support and relevant AI models right into, you know, that goal. And you can see I'm highlighting generative answering being one of those new capabilities to bear. Right? So here we have a a company that is building, gen AI solutions telling you it's not gonna solve all of your problems, because it shouldn't. No technology ever will. Right? The idea is that as you're looking at a website, you wanna give people the opportunity to search, and you wanna help complete their searches with something like query suggest, a behavioral, ML model that we run to you know, you're all experience it with Google. It's like, how do I and it knows what you're already typing before. You're like, that's creepy, but it's also kinda cool because it's exactly what I I wanted to say here. Right? So that that's the idea. You wanna make sure that we are, getting AI to work together as a team along the entire customer journey. So with that, though, we've got to take a look here, at not just a AI, but one of the big risks that comes up to all of our organizations. Right? There's certainly risks around hallucinations, risks around security. Everyone's worried about that, but there there's also a risk, that is of our own making here as service and support leaders. And that's the risk of siloed experiences. Right? And so those of you that are thinking we need to put GenAI into, you know, the mix of tools and capabilities that we have. We need to add Gen AI to, you know, our customers journeys. We've seen this play out before. Right? In a fractured experience. And so if you're adding gen a gen AI, it cannot become a separate siloed interaction channel. Let me show you that the vision of two worlds here, and and see which ones are just feels better to you, or which one feels a little more common at times. This is what I like to call the fractured digital experience. And sadly, this is the way many organizations have invested, over the past ten to fifteen years. Right? We added a search box, and then some other part of the business added a different search box. And then our product team added a third search box, and they're all run by different engines. And then we've got the ask a question box. And then you can start a chat with someone. We've got navigation rules and okay. Now we're gonna add a gen AI bot on top of this. And what happens is customers are gonna get different information from each of these because they're not unified on the back end. Right? There's no one single source of truth that that is driving these interactions forward. And, sadly, this has been the case for us in a lot of ways. And, Devin, you can you can add on top of that that you're gonna find a lot of these different boxes on the same digital property. Like, you can have a search query, a chatbot, and, you know, different facets or whatnot, on one digital property. So let's say your your product documentation page, and then you're gonna find this kinda type of the same experience, let's say, on your support portal, and they're not using the same content repository. So what happens is your customers go through your documentation site first. They go through one experience, and then they go to the support site. They go through a different experience, and they're actually going through cell service hoops before trying to find their answer, which adds a lot of friction by looking at it that way. So, yes, there's different ways to interact with the content, but there's also different digital properties, which is, you know, adds more complexity and friction to the experience. Yeah. More confusion as well. And, you know, one thing that that I I've seen in the past, there's this idea called the uncertainty reduction principle. Right? This was a, a study done in the nineteen seventies by two, the relationships between humans and companies. There is an inherent amount of uncertainty and skepticism. And that uncertainty only builds when there is a lack of transparency into, like, where issues, or where information has come from. There's a lack of proactivity on behalf of a a company or another entity. And one of the best use cases that, that this uncertainty reduction theory has been put in was online dating. Remember when that was a whole brand new thing and people were like, I don't know if I'm gonna go out to dinner with some stranger who I just met on the Internet. Right? Like, now, today, it's like, yep. Swipe swipe swipe. That's what people do to meet each other. But the idea, when it was first, put in was the the more information you share, the more proactive you are about sharing information. The more, you know, pictures and tidbits you share about yourself with someone, the less uncertainty that there is. And that was, principle used, and that's what we're seeing happen here as well. Right? This uncertainty being created. Customers are already doubting themselves in their own ability. We don't need to give them any more reason to do that. To the a couple of questions that that that I see coming into the the chat here. Yeah. It it is about it's not necessarily one knowledge base, but it's one unified index. And we'll show you what that means in just a minute here, right, and how powerful that can be. But it's moving from that world of fractured experience to this world of, you know, a unified AI powered experience. Right? This digital engagement paradigm shift that is happening. We're moving from, you know, a search box or question box or chat box and all of these things to one intent box. Right? Tell us what you're trying to do, and we will help you get there. Right? Like, think of this just visually if it was only this intent box. Right? Where how much, easier that feels to the customer? How much less you're giving them to look at. Alright. This is why Google was so much, more successful than Yahoo in the early days of search. Yahoo had stock score stocks and sports scores and weather and news and all of these things. And customers like, I don't really know where to go. Google said, just tell us what you're trying to do. Right? And this is that natural evolution that you see happening here. Tell us the thing you're trying to do, and we will help to guide you along the way. Right? And so let let's show you now how that becomes a reality. I'm gonna kick it over to to Patrick, who's gonna take you through, you know, exactly how this starts to to happen and how this unified intent box works. Yep. Thanks. Thanks, Devin. So we talked to the the question in the chat around having, like, one knowledge base, this is where the whole thing actually begins. Right? So when you're thinking about, you know, generative answering or Gen AI, I think the term garbage in, garbage out has been, you know, used in many, many shapes and and forms, but I think it's taking its meaning even more when we're thinking about Gen AI because we're actually generating something new based on something else. So if your your content is not factual, if it's not accurate, if it's not up to date, you're gonna be generating something that will be inaccurate, which is not, like, necessarily gonna be a a hallucination, but it could be. So getting having the ability to build that plumbing. So if you build the slide a a little bit, Devin, it really starts with your your content sources and being able to have that connectivity to say, okay. It's not one digital property with one, content source and another one with another, like, we've seen historically. I'll take the example once again of a of a online documentation product documentation where it's, you know, driven by your product management the the support portal, and then the chatbot is workflow based or has some type of conversational AI in there that uses a different corpus of content. What you need to do is bring all of that together and kinda say, okay. What is the most relevant content that the user needs at this particular time? And that's where, you know, a lot of people, minimize the importance of it. But the importance of the plumbing and the connectivity to really leverage all of your enterprise content to be able to leverage this technology to its maximum potential, that is one of the key things that you need to think about and that you need to have. It's it's it's really about the plumbing and the content strategy. Right? If you don't have the right content, you're not gonna generate accurate answers. So that's don't undermine it. A lot of people like like what Devin said at the beginning of the webinar. If we had a button and just put Jenny and I in place and everything was good, that's not gonna happen. You really need to think about this. So that's that's number one. Number two is how you manage all of this. And, definitely, what we've seen have the most success because you have access to all of that corpus of information, is having this this unified hybrid index in which you have all the documents, you have the security, you have the permissions. So you understand what content you have, who has access to it, But you also to be able to drive semantic search and Geni capabilities is you now have the embeddings or, like, kind of chunking the documents into bits and pieces so that, the the the technology understands what that paragraph is about. So when someone asks a question, you can say, okay. This is the paragraph that answers this person's question. So you need to have these embeddings, which are transferred into vectors, which gets into a little bit technicality here. But bringing all of this in one big index makes this whole thing much more efficient because, you know, as we go to the next step, which is number three, is is being able to drive that level of of relevance. And and how you go about retrieving the relevant content, you need to have these these embeddings and and your content all in one place, or else it becomes very cumbersome. You're gonna have different experiences through lexical search versus semantic search. So having it all in one place definitely will drive better results and allow you to drive this level of relevance for your your users or your customers, which we like to refer to as relevant segmented retrieval. So it's not just about one document. It's not just about understanding the intent. It's making sure that you're leveraging everything you have in terms of contextual information and and what you understand about the intent to ground the prompt and engineer the prompt in a way that will drive the most accurate, relevant, and personalized answer to that user. So this this is key. So as you see, we're we're building here plumbing. You need the repository with all the information to be able to drive that level of personalization and and relevance. Now the fourth pillar, excuse me, is is that unification. It's that intent box. Right? What we see and what we hear from our customers is there is going to be a convergence of channels. Right? You you don't wanna have this fractured experience anymore. You don't need a search experience and a chatbot anymore. You bring all this together because that's what customers are gonna expect. They wanna have this one stop shop where they're gonna be able to self serve, and the way to do it is to bring this this combination of capabilities that will leverage large language models, which will allow you to drive these generative answers. So generative answering is definitely a key, but you're not gonna be able to do it alone. You need semantic search capabilities. You need to have these recommendation engines that are relevant to the user that are based on on AI because users and customers have become accustomed and expect now to have this level of of personalization. Right? When you open Netflix, when you open Spotify, you know, you don't necessarily have to search because you have all these recommendations, and people expect that now in in all of their their experiences regardless of the brand, regardless if it's b two c or b two b, these these behaviors have transferred over to the business world, and you need to be able to do that. And by having one intent box, it makes this whole thing much easier, especially if you have the first three, pillars, that builds on top of that. And then the final piece, which is the fifth aspect, is closed loop learning. Right? And this is something that is is key. You need to understand what drives the best results. So looking at the behavioral user interaction data, making sure that you're able to leverage these things to feed the relevance models so that every time something happens and you understand that, okay, someone searched for a, and the end result was b, that was a successful experience, Then the next time someone searches for a, you will boost b even more because you understand that this yields results. So definitely having this closed loop learning process is going to be key to, you know, implement not just Gen AI in your service experience, but to really build the foundation to the to delivering the experiences that your users and your customers expect in today's day and age. Yeah. And, you know, that that's so important. You know, there's like, well, this is how you build great relevant generated answers. It's how you just build great answering and results for your customers for you to stop. Right? This is one solid foundation, upon which we want to build our unified experiences for our customers. And, Patrick, I know that you've taken a look at this in a couple of ways, right, of exactly what this could look like for, you know, a company going through this. Do you mind taking us through this component as well? No. Of course. So we we've talked about the intent box. And what you see here on the screen is is what this unified view would look like for users and and really, you know, leveraging everything that you have in terms of, you know, generated answers, which is the first box you see right there on the right, is, you know, as they search, as they interact, or engage with your content, because that's what they're gonna do now is they're gonna engage with your information. You need to make this, you know, easy to navigate. You need to show them the sources of truth to say that, hey. We're not just saying anything here. Here's what we based ourselves on in terms of of generating this answer for you, And this will move into a more conversational experience as these technologies mature, which is you know, I asked a question, you answered you gave me an answer back. I'm gonna have a follow-up question, or it's possible that I have a follow-up question because I'm working through my problem. So you wanna make that conversational. But at the same time, you don't wanna lose sight of all of this personalization aspect, where if I ask a question and you know that others have asked similar questions or follow-up questions to my questions, You wanna have this this, you know, people also ask, which we're starting to see a lot in in different experiences, you know, like like the Google experience and so on and so forth. So this is this is gonna be another important aspect of of when you're thinking of delivering these complete experiences. You know, these two kinda go hand in hand along with, you know, keeping the the ability for users to to navigate through a list of results. It's great that they have an answer, but they want they might wanna know a little bit more. They They might wanna actually read that thirty page PDF that you have because they they need it. They need it in their work. They need it in in whatever they're doing. So only keep that there and make sure that you have that recommendations engine as well, which means, okay. You you're looking for this type of information. We know that others who looked for this also consumed x y z or bought a b c product or whatnot. These recommendations are are what people are now accustomed to. And finally, on the left hand side, you know, we we talked a lot about engagement engaging with your content, having this dynamic experience where people can can filter through the content, make it more precise, get better recommendations is gonna be key, and that can that can't be static. Right? You're you're not gonna wanna list of thirty different facets from people to to, to filter on. They're probably not gonna use it. It's it's too much of a hassle. It's too complex. So you wanna make sure that those dynamic those facets are dynamic. They are, auto ranking based on what they're searching for and your purpose of of knowledge or or documents. So all this is part of a a complete experience, but I'll go back to what we mentioned around the unification of the experience. If if you're going about this and saying, okay. We're gonna put this experience on our documentation page, and we're gonna put this experience on our support portal, and we're gonna put this, you know, somewhere else and all the different channels that you have. Yes. You're gonna have a a very, pleasant experience for your users. But if you're not leveraging the content from all these places so that it's a unified experience, which means whether I go to the doc site, whether I go to the website, whether I go to the support portal, if I ask the same question, I should have the same results based on the intent and what you know about me. Definitely, if I'm on a website, you're probably gonna, you know, wanna make it more salesy, more marketing because that's the purpose of the website versus a funnel support portal. You might wanna think about it more from a a troubleshooting perspective or how to questions, which might change a little bit the relevance because of the intent. But it shouldn't have these these different digital properties should not leverage different pieces of content. They should leverage the same content, and the experience should be unified across the board so that your customers no longer need to jump through different hoops or go through different sites to have access to all the information that they should have access to regardless of the landing place or their starting point in their digital journey. And that that's probably the the main takeaway that I would say that, you know, you you probably wanna have from this webinar is is how are we gonna unify this? How are we gonna break down the silos between our digital properties and even our assistant support leveraging this technology? Because that's key. You wanna remove the friction. You wanna make it easy. You know, CSAT is definitely something that we look at from a self-service perspective. I encourage you to start looking at customer effort score and and and how easy you're making it for your customers to resolve their issues from a self-service perspective because that's that's really actionable, and it it can easily help you identify where those friction points are across that digital journey. Mhmm. And, you know, to to me, Patrick, this is so much, not just about the issue I'm solving today, but what comes next? How do we continue to build that that engagement layer to, you know, guide our customers along that journey with us? That that's something, I've heard from a couple of Coveo customers that that I've been speaking with recently. It's like, we're putting in more purposeful digital guidance to say, you know, sometimes we can't answer this question, with with a a generative answer or a document or a search result. You gotta talk to a human for that, and that's okay. Customers are fine with that as long as that that answer, is something that helps them to get toward, again, their ultimate goal of resolution. So what what does this start to look like? When should we start to use GenAI? Let let me, you know, quickly share a couple of things here. We'll have plenty of time for q and a at the end. And we'll, end with a conversation around one of our clients, Zero, who's been doing this well. But, like, when do we even wanna interject generative answering? Well, what about a question like this? Well, what time is the Logan Square branch? Imagine we're a financial services organization, is is open on Milwaukee Avenue on Saturdays. That's my neighborhood here in Chicago, by the way. Well, if that's the question, generating an answer, too much club. Right? Just give them a result through search. Here are the hours for the branch. Right? That's all I need to know. That's the bit of information. But as Patrick was talking about, you don't just leave them there. Right? Here's a search result. Off you go. If you know something about your customer, you wanna recommend the question they haven't thought to ask you yet. Right? So if someone how do I save for college? Of course, you're gonna show them, information about starting a five two nine, but they haven't thought about wait. What do I do to set up a home budget? Right? So you pop that as a recommendation knowing, as Rex said, what other people have asked and what's been successful in that situation. Here's the next thing that might come down the line. Right? So you wanna continue that journey. But when customers have either more complex questions, questions that are multipart, or questions that would require a lot of effort on their part to read. Here's a seventy five page document, and you're probably only gonna need pages, you know, thirty two to thirty nine. I can't go looking for those things. Right? When it's too much effort for me as a customer. I like to think of these as level two customer questions. There's not a named and known answer directly in your knowledge base, but the answer is there. It exists, and you can use the generative capability in order to do that. Right? So what's the difference between a personal loan and a commercial loan? Hey. Good question. Here's a generated answer for you. Along the bottom, here are the three documents that that came from. If you need to if you read this, you're like, actually, you're right. It is a personal loan that I need. Boom. Personal loans. But number three. Or I might need to have a follow-up question. Right? Again, as far as we we see all of these things starting to converge in the way that that they're working. And, that's where the the world is gonna go. Right? I I wanna talk to your company's information digitally first, and then maybe with the human, next. And so we we've seen one company, you know, called Xero. They're in a software firm out of New Zealand. They make accounting software. And, Patrick, I know you've been working with them, you know, pretty closely here and and talk to them, a lot about not just how they're using Govea, but how they're using generative AI in particular. You mind telling us a bit about what they're doing? Yeah. Well, you know, Xero was one of our our first, customers to leverage this this, technology as part of our early adopter program. And, they were they were kind enough to test this along with us and and help design it. But one of the things that, I think is is key if you wanna see what type of impact you can have, by by doing this around not just, you know, the cost reduction because, you know, there there is that that twenty percent increase in self-service resolution, which clearly has an impact on on your your your costs. You're lowering your you're lowering your cost to serve, while, you know, adding stickiness to your customer experience. But I think the key thing here, if you look at Nigel's quote, is is definitely that stickiness to their platform, which means every time they remove friction from the experience, You know? Customers are more willing to stick around because they're making it easy. And I wanna highlight that that, you know, forty percent reduction in average time to resolution and, time spent searching. And that that's key when you're thinking about improving the customer experience, removing the friction from the experience. And at the same time, as you are generating accurate answers, the the the confidence of your customers dramatically increases in your your ability to serve them in the way that they want to be served. Mhmm. And this is just an example for self-service, but it's definitely speaking around the impacts that this type of technology can have when you leverage, you know, all the the five pillars that we touched on a little bit earlier, to make sure that it is secure, that you do level leverage the the level of personalization that your customers are accustomed to, you can have significant impact in a very short amount of time. And this is what Xero is able to achieve by by doing this, by doing exactly what we talked about today. That that that's fantastic. I know they're just getting started, on their journey as well. Right? They they saw this in the first six weeks, and they're continuing to to learn and mature, and that's what makes it so much better. Let's move into the the q and a here because I see that there, you know, are a couple of things. And by the way, for those of you interested in, you know, continuing these conversations, as I said, we're we're happy to have them. You should also, look to join us in what we call relevance three sixty. This QR code will get you there, right where we've got folks from, you know, Forrester, Zach Cass, who came from OpenAI, speaking at this event, you know, to continue these types of conversations. Chloe, I I see there's some, pretty cool questions coming in for anyone, like, warm up your fingers, put it in the the chat or the the q and a here. We wanna take these from the the top. Yeah. Absolutely. So thanks so much, guys, for that presentation. That was really interesting. Thanks for everyone who submitted questions so far. You do still have time, if you wanna submit questions for Devon and Patrick. But for now, Guy asked considering the speed and rise of the AI tech tide with currents coming from more and more vendor angles, as as you mentioned earlier, Devon. Do you see the need for new positions being onboarded for those focused on that specific balance of human agent enhancement with your tech applications. I've got a thought on this for, you know, one type of I'm not sure if it's a new role, but it might be a readjustment of an existing role when it comes to generative AI. And then, Patrick, I don't know if you have any thoughts as well, but one is QA. Right? Any of us that have been in the contact center for a long time know QA is such an important function. And I wonder if anyone's thought, how are we gonna QA generated answers? And and the good news is it's gonna be kind of in the same way. Right? You're gonna put, some humans against these generated answers to look for accuracy, look for tone, look for, some of the those same components that that we've had. So to me, I see that being, you know, the the type of role that we would have to adjust. We might have to spend less time focused on some of our humans and more time a little bit focused on that, or at least including GenAI as a part of our our normal QA processes. Patrick, any other, roles or new positions that that you've run into or heard of? I'll go along the same lines as what you mentioned, which is the the evolution of the agent role Yeah. And and support models. I'll push it even all the way to support models. You know, we're very in tech at least, very familiar with the tiered support model where we have level one for the easy stuff, level two, level three, and things escalate if one level cannot, cannot answer the questions. In in financial services, it's it's one department versus another, and it's very, very segmented into who does what. And and sometimes, you know, you get as a customer, you get transferred from one place to another, and it becomes a very frustrating experience because of that. I think with GenAI and self-service and the impacts that it will have, like we saw, the zero use case or case study, is that things that are coming into the support organization will shift as we mentioned earlier. So your support model around the tiered approach might move to a more collaborative approach where, you know, there might not be a need so much for l one, but your l two and l three will become much more, involved in the resolution, wanna collaborate together, wanna collaborate with SMEs. And I think the role of the agent will transform more into a a knowledge worker, which is the goal is to document things so that, you know, GenAI will be able to leverage that piece of content and generate answers for the next customer that encounters this issue. So, yes, you might see it in a, you know, a an increase in in case deflection or self-service success by twenty percent. I wouldn't go all the way to say, well, we're gonna cut twenty percent of our workforce because you want to leverage these people to drive your knowledge management and your content strategy. So I think there will be an evolution of the role and being able to leverage their skill set around how they understand your products and services to be able to create and maintain that corpus of knowledge and content for it to be actual, factual, and and be able to generate accurate answers. And then you QA the answers, right, to make sure that they are accurate and you identify content gaps, and and you continue building, and you come you you now have this continuous improvement loop around that, which I think is gonna be key. Thank you both. And I do have another question here, from an anonymous attendee which says, is this a Coveo product developed from scratch, or is it powered by another GenAI platform such as OpenAI or Copilot? The, we leverage the entire Coveo platform, everything you saw in those five pillars is Coveo built. The only component that is not Coveo built is the large language model. So right now, we are using, GPT three point five, as our large language model, but we're leveraging the entire Coveo capabilities to be able to identify the the snippets that are the most relevant across your entire, you know, knowledge base or or content enterprise content. And we actually engineer the prompt and ask the large language model to generate an answer solely based on this content and not the information it was trained on. So we're really leveraging the LLM for its its language generation capabilities, not the information that it was trained on to do it, which allows us to reduce the risk significantly of any type of hallucination as well as, making sure that the answers are accurate as long as your content is is, is accurate as well. So it is a mix, but the majority of everything that we talked about is Coveo platform besides the large language model. Yeah. The the way that I think of it is we're using GPT like an English professor. Right? Here's a large prompt that, you know, that they GPT doesn't see your, your company's content. It's vectorized before it goes over. Right? But it's like, make this better. Make make this smarter. You know, be an English professor, and then help me to write this in a better way. That's the the way to use it. And, again, as Patrick said, the most important thing is finding the right pieces of content to send over. If you don't get that and you're using, you know, OpenAI as a fact base, it was never built to give correct answers. It was built to transform, prompts into answers. So that's something to consider, and you wanna generate that from your own company's information. Thanks, guys, for those answers, and thanks for everyone that submitted questions. There are a couple that we didn't get to, but we are unfortunately out of time. So one thing that I will just let everyone know is that this has been recorded. As I mentioned earlier, you will be able to access this on demand. You'll be getting an email from CX network, and you can also do this via the Zoom lobby and via the CX network website. Another thing to be aware of is that Devon and Patrick, both have their LinkedIns connected to their Zoom profiles. So if you did wanna take this conversation offline, I'm sure that they will be happy to oblige. Yep. And I saw that there was an addendum to that last question. Let's follow-up offline, with Aki. Awesome. So, that's all from us today. Thank you so much for joining us wherever you've joined us from. Don't forget to check out the resources box via the lobby, and hopefully, we'll see you next time. Thanks so much, everyone. Thank you. Thank you.