Audience, our next session is brought to you by our partners over at Coveo, and will be presented by Devon Poole, who's on the line with us now. Devon is a senior marketing manager at Coveo, and he's been researching and advising on CX, for the last fifteen years. He's previously worked at CEB and Gartner. So you're in great hands. He knows his stuff. But before I do hand over, I just wanted to remind everyone of a few housekeeping points. So first of all, please remember that our sessions today and yesterday as well are all being recorded. So if you do want to revisit any talking points, you can go ahead and do that via the Zoom lobby. Next, please remember that our resources section is open and we've got free downloadable content from our partners to look at. And finally, our q and a box is open. Any questions that you have for Devon, or our speakers later on, please do pop them in, the box, and we will get back to you, with as many as we can. But for now, I'm going to hand over to you, Devon. Thank you so much for joining us. Yes. Thank you, Chloe, and please ask questions. I I love questions. Those are the ways that we know folks are engaged with this. So, you know, with that, I I know we we, have about twenty eight minutes here, and let's hop right in. You know, we we had initially closed, and we we talked to, you know, so your members are talking so much about the ability of omnichannel systems these days. And as you said, I've been studying the customer support space for years. When it was back it was the this web thing was a new thing, and now, you know, the the web is the the primary driver. But as we're gonna talk about today, it is certainly not the only driver, of resolution. Right? And when we step back and remind ourselves, what are we all in this for? We are here to solve the problems that our customers have with the products that that our organization sell or the services we provide. Right. And, over time, while customer preferences have shifted, they've shifted largely because provide that enables our customers, has shifted and evolved as well. You know? But one thing has always and will always remain constant. Customers want you to solve their problems quickly and easily the first time. Right? That's the goal for all of us. And, you know, over time, though, the the definition of, you know, easy has pretty much meant the the same thing, right, quickly meant the same thing. But, you know, first time is not the same. And that's why what we're gonna talk about here is the idea of the connected customer journey and how, you know, this little thing that we've all probably heard of these days, AI, and not just the generative type of AI, but AI can help us to solve those problems, help us to create that connected journey for our customers. Right? So let's, as always, start with a a little data, a little bit of level setting for us to say, hey. Where are we today? And we know this to be true from our conversations, from, you know, research out there that customers to make they they do demand self-service. They they want self-service. That's the first place that they tend to go. Just looking left to right on the page here, you know, seventy percent of customers today say that they prefer prefer self-service over contacting a CSR or a rep, some sort of live channel. Right? In the middle from CX network, eighty one percent of customers say they try to solve their problems themselves before using assisted channels. Right? Like, we go out here first. And on the right hand side, thirty nine percent. You know, that that's a huge number of customers when you read this stat. This comes from a Gartner study. This is thirty nine percent of Gen z customers who say that they'll give up on an issue entirely if there's no way for them to self serve. Now, yes, me, sounds a little crazy, but it's not just Gen z here. Right? This is also, you know, thirty one percent of millennials and twenty eight percent of Gen x say the same things. So it's not just, oh, these kids are, you know, complaining. No. It's becoming a real thing that for certain issues, right, for certain ways that the people seek service, this is what they're looking for. Right? I just wanna be able to self serve on those things that I deem so simple, so easy. Now what we know is for the vast majority of things, they start in self-service. They don't resolve there. Only thirteen percent of, service journeys end after the first contact in a self-service channel. Right? In fact, the average customer and this comes from some Salesforce data, their, state of the connected customer the this year. The average customer uses eight to ten channels throughout their life cycle. Now that's not any one journey. Right? That that's not, oh, anytime I contact the company with a problem, I'm gonna use eight to ten channels. But what this tells us and what, you know, we we need to think about here is your customers are not channel centric. Your customers are not using only one and one type of channel for everything. Customers realize that different channels have different value positions. Customers realize that, you know, let's give them a little credit because often we think of customers, we think of our worst customers and, like, who amongst us now that we're in the business of customer service haven't become our own worst customers. Right? But, the the customers, they they look at the menu, the buffet of channels. Right? And they say, yeah. I'll try all of that, please. Right? I'm gonna use what you have on offer for me. And that in and of itself is, you know, sort of the foundation of this omnichannel promise that we've all been looking at, that we've all been talking about. If only we can get customers to, you know, use these new channels and try them, well, that that's what they're doing. Right? Your customers, though, they see you as this omnichannel world. You've been pitching it. They're saying, yeah. I want one channel. But and, of course, you you probably knew there'd be a but in this. There is a problem. Right? The this omnichannel experience, it is seriously lacking context in the journey that the customer is providing, or the customer is being provided. Right? So, look left to right on this page. It's from Forrester. They found nearly two thirds of customers found that when they had to switch channels. Right? I Coveo from, you know, the the web or a chat to a phone call or, you know, some sort of transition here, two out of three say, yeah. I've gotta reexplain my issue. That's that that that's a tough one for customers to stomach. Right? Nearly three quarters of customer service journeys touch at least two channels. Right? So customers say they have to reexplain things, and it's happening all the time. And seventy one percent this is sort of the the silver lining here from that same Salesforce study. Seventy one percent of customers say that they prefer a different channel depending on the context of their situation, right, depending on what they're trying to solve. So that's good. Hey. At least the customer is trying to figure out what the best channel is. But the the killer here, right, when a channel switch happens, and it happens three out of four times, right, those channel transitions more often than not are deemed high effort for customers. And those of you that have been following this space for, you know, the the twelve, fifteen years now that that the world, has shifted toward the effortless experience and ease of resolution, choose you your term. Right? Know that high effort is a customer loyalty killer. It's a cost killer. It is a experience killer. Right? So when things are high effort and channel switches are happening and they are high effort, that is bad. So what can we say do about it? Well, it will information. It is due to information that, the customer is looking for and cannot easily find, or the information that they're finding is not relevant to their situation. Right? So what we look at and one of the things that that we talk about here at at Coveo, right, are are about relevance. Relevance augmented experience. And that starts with your content. Right? Tiny little thing that we're gonna build up into here, but it starts with where do you keep the information that your customers need to solve problems? Right? Some of that could be videos, on YouTube. You may keep things in confluence. You may keep things in knowledge bases. You may keep things on your website. You the point is you probably keep things in, on average for us, what we see somewhere between eight to twelve, content repositories being indexed into Coveo right now. Right? It's structured. It's unstructured. And then you need contextualizing data sources from, you know, your CRM or CRMs, plural, depending on how many you have, your CCaaS system, your, you know, different knowledge systems. So wherever you keep information about that customer to contextualize that experience, who is this person? What products do they hold? What information do they have access to? Right? That's where all of this starts. Gathering those things, not moving them, not lifting and shifting them, but connecting them. Connecting them securely into a single spot. Right? Drawing from out of the box here secure connectors that can look at your information, that that can read your information, that can, you know, do what we call chunking. And I cannot say that word without thinking about the Goonies. Right, looking at those relevant chunks, generating vectors so that you can keep information secure, and pulling that all into an indexing pipeline saying, now we know how to find this information wherever it is. It's securely connected in and can move into a unified hybrid index, which can then be enriched. Right? This is, again, most of us are used to dealing with Google. Google can find anything you need, these days. And, and when it's your own unified hybrid index, you can boost certain things. You can pull things back. You can help, you know, ensure what people see, but that in rec that that index index is enriched, with, you know, embeddings and vectors contained within it. Right? So moving from securing, securely connecting your content to, putting into the hybrid index to be able to find that content to retrieving it, right, at time, of query from a customer. Right? And who amongst us when you go to, you know, find information, you search for it. I need to look for it. Give me a place to tell you what I need. In fact, one of our clients who's a large national telco who you definitely know, said, look. If we wanna know anything, about our customers and what they want, what they're looking for, what they're trying to find, we put a search box in front of them because the search box, customers don't lie to it. Right? Customers don't try to outguess it. They just put in what I need. Tell me what you're trying to find, and we can retrieve that information. We can throw, you know, relevance AI models. So there's behavioral models. There's large language models. There's, different ways that you can find that and find the right content. You ground that content in a fact base, not just for large language model generation, but for search results, for recommendations to say, what has been useful for customers in this situation before? What do you know that customers have done that has been successful? And then you have the context that that comes into that. So, again, building this pipeline. You're retrieving the right information, and then you're putting it in front of the customer in a unified way. One single box, you know, and we'll come back to to this idea in a minute. It doesn't need to go through an LLM, but it certainly can go through an LLM to generate an answer, but not all things need to generate an answer. Sometimes just a search result. We have a document that answers this or an FAQ or a thing or, you know, some a video for a customer to watch that answers this. Just go to that. You don't need to generate an answer. Sometimes, sound of anyone's a golf, advocate, but sometimes that's what's called too much club. Right? But, certainly, it's there. You want that to be an option. But there's also the question the customer didn't know to ask, and that's where recommendations come in. And, of course, we want to personalize that information with the context that we have about that customer along the way. It doesn't stop there, though. Every click, every search, every input, every action that the customer has can be fed right back into your own retrieval, platform to say what worked, what didn't, so that the machine learning models are constantly getting smarter. Right? Usage analytics, usage context, and behavior is directed right back into that retrieval of what worked, what didn't. Because I don't know about you all, but, we don't live, at Coveo in a static business environment where things aren't changing frequently. And so everything needs to be constantly updated. So this is how we're starting to create that unified relevant experience for customers, not just digitally, but as they move through the world, as they move through that service journey. So let me start to show you here what that looks like. How does that feel, and what companies are already doing that today? So, these are the keys that that we are looking at here to improve that, you know, omnichannel service experience. It's about creating those unified digital experiences. Nothing says call me faster than a customer who gets two different answers from two different parts of your website to the same question. Right? So it starts digital. We know that. And there is a huge component of we wanna contain the calls that we wanna contain the contacts that we can hear. Right? Shift left. Get those contacts, not just in a website, but if you're a, you know, SaaS company or you have an app where your customers interact with you, contain questions there as well. Solve problems where the customer's already working, but you're never gonna be able to solve everything. Right? You'd, that you're gonna have a space for that. So you've got to connect that journey through, and then you've gotta make sure that when someone reaches a human, that human is knowledgeable knowledgeable about the customer themselves, knowledgeable about the solutions that exist to their problems, and knowledgeable about the journey that customer took to get there. So let's start with these unified digital experiences. I I said I'd come back to that box number four we saw a minute ago. Right? What we love to call the intent box. What's happened largely, over the past ten years, and it's, of sort of no one's fault at the organization. We've invested in solutions for our customers, and we've invested in a fractured manner. Right? We've got several different search boxes depending on what part of our website we may be talking about. You might have two or three different chatbot solutions. You've got UX design in there. You've got navigation menus. You've got you name it. And we've invested in these things in ways that seem to be siloed. Right? It's the same problem that we lifted and shifted over from our contact center years ago of siloed information, siloed queues, siloed agents, and largely that that's reached the digital world. And now is the time that, organizations are starting to make that investment in this unified digital experience, at least based on the hundreds of enterprises that we are talking to day in, day out. This is the thought process. And if we can give someone one intent box, tell me what you're trying to do. Right? And that box can be placed in a mobile app. That box can be placed in some sort of, you know, chat widget. That box can be placed in just a search box on the page. You click get help. I've long advised companies when, you know, you you want to go to your marketing team who owns your website and say, hey. We need some stuff for support. In fact, what we need is less. We need one spot in a prominent position that says, tell me you need help. Cool. I'm gonna direct you to a help page and put an intent box in front of them. That intent box can, if you move in the blue boxes down the middle of the page, it it can provide, a generated answer. Right? It can provide an abstract if that's what's required. It can also, you know, give you that next question to that. Oh, here's a follow-up question, whether it's, hey. Converse with the large language model, or here's the question you didn't know to ask, or people also ask, or here are suggestion follow-up questions for you. It can help direct you to a human conversation. It can give you search results. Right, and those can be elevated to the top when, hey. There's a document that solves this problem for you. Right? It can be recommendations. Hey. Here's the thing you didn't even think about yet. Right? It's may not even be related. If you're asking about college savings and we're recommending a five two nine, But, hey. You know, you probably would have never thought about our home budgeting tool if we're a, you know, college savings plan or a retail bank. You should fill out this home budgeting tool. Ten minutes easy, and it can help you. Right? Like, those types of experiences. So what's happening here is, you know, if you look at the, green text on the right, you're drawing from a single source of truth. You're, using AI and ML models to personalize that experience based on all of that contextualized information. You're guiding the customer through that experience. Here's what you need, one place for help, and we got you, and you're providing that more conversational interface. In fact, let's take a look at a company called Xero. Xero is a, software company out of New Zealand that we've worked with for many years. They've got many different, Coveo deployments. Most recently, taking our, Coveo generative solutions model, right, and putting that out in front of their customers, exposing it. They've been doing it for, you know, six months now. And in the first six weeks of this, they found they were able to improve self-service containment by twenty percent. Now why did that happen? Well, they attribute that to a couple of poor things that you see along the left hand side of the page here. First, they're providing faster answers. They're not making their customers read long accounting software documentation. Right? You can imagine, the the documentation there is hard, and they are able to generate synthesized, you know, short answers. Whether it's, hey. You might have had to go to three different documents to read this. That happens quite often. You know? It's like the answer was there in self-service, but it's kind of a PETA for us to go get it. Man, we can synthesize that really quickly. And most importantly, in in number two there, it's secure, so it knows who the customer is, and it provides them only with documentation they have access to, and it's accurate. And I've just seen some data that shows, man, hallucinations are coming back in a in a big time. Well, for us, it's, you know, over ninety four percent accurate that we're seeing here. And, man, that that is fantastic, to see because, of that approach that I showed you of grounding, that answer in your own fact base, applying those very mature relevance AI models. That's where the accuracy is delivered from. And they say that, you know, their customers are confident in these answers because they're providing the citations. You're starting to see that pop up. That was, you know, the first thing we we built into our model was, here's the citations and the accompanying links. Yeah. Here's where this answer came from. And if you wanna read more, go for it. This is where oh, okay. So I know this came out of your own documentation, and I can read a little more if I have to, but I've got a good gist of what's going on here and might need to do more research. But what they found in the, diamond on the bottom, middle there that they found about forty percent reduction in time to resolution, meaning customers were spending about forty percent less time on the page, and that's huge. What that means is customers are getting those answers that they need. Right? So we've got tons of of, you know, videos, and stuff about the folks from Xero talking about what they've done on our site. Go visit those if if you want more information on how this worked. But, man, fantastic result to provide that experience because, again, they're still providing search results. They're still providing recommendations to their customers as well, and they're adding on this generative model to do that. The next is about connecting journeys, and I apologize for a bit of an eye chart here. Right? But, let's start at the colored boxes along the top here, and how you deploy AI along the entire service journey, starting left to right from, you know, public access websites. We wanna provide a lot of service information there, things that anyone can have access to that could solve a problem for a customer. So our customers want easy access to information. Moving to, you know, if I'm logged into some sort of product, that could be an app. Again, that could be a SaaS product. That could be, some other product experience. You put, you know, contextualize, knowledge there to self-service all the way through to the contact center. Following that green bar along the middle, you're gathering the context of that journey along the way and applying different user interfaces from search to recommendations to generate an generative answering and conversations to, you know, relevance wrapping all of that in a filter. So you can just see some of the models along the bottom here that, you know, are composable, and you can say, man, we're we're gonna, you know, play with this. And in our in our website, we're focused on, you know, not providing generative answering, but we wanna use bottom left large language models to do that. Well, there's something called smart snippets that just finds that relevant chunk, that relevant passage in a document. You know, maybe you're a financial institution, and, you have a forty page document, and you only need one paragraph that's on page thirty six to answer that customer's question, well, that just pops right up for the customer there, and that's the paragraph they see. And, okay, that's the answer. You don't need to go any further. You don't need to generate an answer. This is an answer in a document. You find it faster. And you can apply these models, you know, along that entire service journey while gathering context for the customer. And that's exactly, what a health care company called Athenahealth has done, you know, for the hundred and forty thousand clinicians that they support and the four hundred agents that help to support them as well as the website. You know, they started by providing more of an intelligent self service model. Right? Using, you know, what we've been looking at so far to build a one stop shop website for their customers. You can sort of see a mock up of it here on the right that just how can we help you? Right? This is a search bar. Tell us what you're trying to do, and we'll help you there. And their customers are getting better results. What they're seeing is twenty six thousand, reduction in quarterly case volume of twenty six thousand cases. That's a lot for for them. Right? Man, they're seeing that over and over and over. What they found, though, is that it wasn't just something they provided for the customers. They flowed through all the way to their agents. That helps to reduce onboarding time, but give the agents the information they need. And, interestingly, you look at the, bottom right, hundred percent of their agents adopted it, and they saw a fifty percent reduction in, calls flowing to their tier three teams because tier one agents just had better information at their fingertips right in their workflow. Right? This flowed into a CRM or a CCaaS or wherever your agents are working, you can put that power of the Coveo model directly into it. And, of course, they got analytics on the back end that help to improve their content strategy, their coaching strategy, their knowledge management strategy. One last component here looking directly at the agents. Right? What we know like, the the headline here says it already there are challenges in the agent experience. Agents can't find what they're looking for. Agents are dealing with customers who have already been online, and they say that they're not getting the right information from their systems and tools, which is why, the agent experience has become a top priority for contact center leaders in twenty twenty four. And it's about providing those agents with powerful tools within their work flows here. Right? These are all, you know, three examples of things that that, can sit in an agent workflow, a generated answer. Hey. There's a difficult conversation that's going on, and I might have to review three, four different documents to do it. Same thing your customers would do. But this is a more complex thing. You can generate an answer, and that can give an agent a head start. In the middle, you can show them, hey. That whole journey that that customer took to get here, we can show you what this customer was doing through user actions. Well, these are the searches they performed. These are the documents they clicked on. This is when they did it. Wow. You've been online searching four times in the last four days. Look, I I don't wanna make you repeat yourself. I can see you've been looking up these things. What questions can I answer about that? Oh, wow. What a different experience. Right? And then just giving them good old power of search, relevant search results. Hey. I need to look for things as this conversation is going on and progressing. That's what you wanna start to do. And that's exactly, what we saw from a Fortune one hundred financial as here. Right, what they were providing were in workflow tools for their agents that help them to reduce their their search to click time by eighty six percent. I'm getting the result first because they didn't have to measure multiple thousand jump words in order to find the right knowledge. That reduced their average handle time by, about eleven percent. And then percent of their answers because they're using that smart snippets model didn't even require a click. Right? They were able to say, oh, yep. Here's the one paragraph I need. I can answer this question. Off you go. So with that, let me check the the q and a box here, because I I think we've got a a couple of of questions coming in. Thank you. Very interested in using AI to sell innovation to a skeptical and conservative audience. How do I engage? That's a great question. Start let let's have a conversation offline there, and talk about that because, you know, what you do is you start small. You start in a contained area, and you show some success. I would, you know, say, let's have a conversation about where the best place would be to start, right, and say, man, we can do this. Maybe we start with an internal audience, and we, show the successes there, and we significantly ramp up from there. And it's not about just generative AI. That's what I'd say. You know, your audience is skeptical of, oh, chat, cheap, b t. We can start elsewhere. There's a lot of cool things with AI you can do that don't involve generative at all, and that's where you may wanna start and and prove your case. With that, we are right at the top of the hour. Chloe, thank you so much for having me. Thank you so much, Devin, and thank you to everyone who joined the session. That was really interesting, and we appreciate you joining us today, Devin. Fantastic. Find me on LinkedIn. We'll talk about this all day long. Have a good one.