Hi. I'm James McGuire, and on today's video, We're talking about artificial intelligence, and we're taking a look at the AI foundations that power what's known as generative answering. And we'll talk about what that is. It's a really interesting topic. To discuss that, I'm joined by an industry expert with me as Juanita Oguin, Senior Director of Product Marketing Platform Solutions at Coveo. We'll need a very good to have you with us today. Thank you, James. I'm really excited to be here to talk about a very popular topic. Indeed. Well, can you tell us a bit about, what what does Coveo do before we get really started? And I know there's a lot people who are familiar with the company, but what does Coveo do for its clients? Yeah. I I'd be happy to. Koveo is a tech company. We've been around since two thousand and five really trying to help enterprises solve for their toughest information management challenges. We started, really focused on unified search. So being able to tap into various systems to surface the most relevant information And since two thousand and five, we've really innovated along the lines of analytics, NLP, AI, of course, which we'll talk about in word today, and really have over seven hundred different customers, some of the largest enterprises that you can think of, that use our technology to power their experiences both internally for employees as well as on their website or digital experiences. Interesting because, you know, I hear so much about, AI and search these days, particularly in the area of generative AI. So it's a It's a very current topic. Alright. Well, why don't you go ahead and start with your presentation, please? It is. Thank you. Happy to jump in. And given it's such an, popular topic, I thought it would be nice to start with some trending headlines. So I'll give everyone just a minute to take a look at these headlines. And for those of you that are working in and testing GenI today, which a number of enterprises are, These headlines may not be surprising. They're around whether your organization is data ready. The old saying garbage in garbage out as it relates to college management and information. Right. So a lot of headlines these days are really focused on the need for quality data in order for AI and gen AI to work. So it feels like we've almost come full circle back to those data and information management challenges. You know, there's interesting point about the the importance of quality data and that we see so many headlines about AI. We don't see quite as many about the importance of of data because, of course, AI goes nowhere without really great data. So it's a it's a key point. Absolutely. What I wanted to do was kinda double down on that a little bit more and to dive deeper. What I find what I found and find most interesting about kind of the rhetoric these days around AI and Jenny is it is so data focused and what we see happening is every major tech vendor trying to raise to find an answer to that data management challenges. And we see things like data lakes being proposed or created by technology companies, we see data warehouses or data clouds start to be promoted rag, which is retrieval augmented generation, emerging in helping AIB more effective specifically generative AI and of course knowledge graphs. And, you know, each one of these is very important. Any per data leak especially if you are processing high volume, have a lot of raw data. That that's important to to leverage. Data, wherever houses and clouds they add a little more structure to that raw data, making it a little bit more, usable. And, of course, important again for companies dealing with terabytes of this information. And RAG also a great way to try and bring semantic search and vectors into establishing relationships with information to find the relevant stuff that you should be to the AI. And then we also see knowledge graphs being promoted and knowledge graphs, of course, I've been around for a while. And I would say they're more explicit versus implicit ways of really identifying what people are looking for. So all of these and each one of them is absolutely critical, but I think the challenge is that these are all really thought about as point solutions independently or being built in siloed environments, and really companies need to think a little bit more broader and holistically about it. The other view I thought was important to introduce into the conversation was that it's not just all about data. You know, data is a is an easy, common, familiar word for us all, but I wanted to bring in an older model, the d I k w model, which stands for data, information, knowledge, and wisdom. And this is, you know, an old organization systems thinking pyramid for helping people understand how you take data turn it into information, that then becomes knowledge, which is wisdom and helping people, you know, be more proactive or project what's you know, what's gonna happen with all of the information that they're processing on a daily basis. And of course, we all go through this as, you know, on a daily basis in terms of trying to understand the pace of business and change, make sense of the information that's coming our way. Really, we are all knowledge workers working in a knowledge economy. Right? And so I think it's important to revisit this model and to understand that in between each one of these, there really is that experience, that judgment, that need to understand context and intent. And of course, there are AI models that are now emerging to help us kind of take the next step you know, in between, each of these layers of of processing information, but you're always gonna need people. You're always gonna need humans or as they say humans in the loop to really process this effectively. So I just wanna point out, you know, data is absolutely important, but it's beyond data that we really need to be thinking about as well. Good point. And one, you know, one team and one function that is starting to emerge as part of these big questions around the quality of information that you're feeding into your models is the knowledge management team. So, I've talked to and worked with a number of knowledge managers across some of the largest enterprises that you can think of. And I find especially pre chat GPT pre gen AI. They were coming up, into some interesting challenges around not getting enough management priority or prioritization on their knowledge management projects, which I found interesting, and also that has now all been, you know, ended with with generative AI because, again, garbage in garbage out. So if you don't have good content and knowledge management practices, your AI and Eugen AI is just not gonna be effective. So what I've talked to a number of them about is understanding knowledge management is absolutely critical. But also being real realistic with the fact that knowledge management cannot be done at scale without AI or AI search specifically. And so that's what you see here on the screen, James, and for those of you joining us today are Some of the challenges that, kinda point out why knowledge managers and enterprises need AI. So if I could just walk through some of these, and James to feel free to stop me. I ain't quite for any clarity. One one thing I think is really interesting that the the the points that we owe people resources are limited, and then on the other the other box as people are needed in AI search augments and scales is that I think some workers are sometimes a little concerned like will will AI put inside a business? Certainly not. I think people are needed more than ever, in fact, knowledgeable people, experienced people who know what they're doing are are are in high demand at this point. Absolutely. And, some of the other points in these is, hopefully, get that across as well. For example, we know knowledge management teams are limited in nature, they're often more horizontal roles support roles supporting different departmental functions. We know that enterprises struggle with managing and making sense of information. And we also know that knowledge managers don't always have the domain expertise because they're seen as more horizontal players, so they're bringing best practices, processes, systems, and tools But the fact of the matter is knowledge management or care leaders are absolutely needed to help drive standards that ways for how you process, store, and maybe even retrieve information, and these knowledge managers are gonna be working hand in hand with those line of business subject matter experts. For example, you could have a, you know, kilometers meter or kilometers practice working hand in hand with your sales department. And, obviously, they're the knowledge management person is not gonna know all of the sales terminology or that domain expertise, but that's why it's a big partnership with the the departmental subject matter expert, the knowledge manager, and then the technology which helps connect all of this and scale in ways that, you know, people can't process terabytes of information on our own, unfortunately. I wanted to go a little bit deeper into, why AI search, and why is this the right path or the right set of technology to even consider. Why surge in this big GenAI AI kind of dialogue and and and, rhetoric. And what we started to see happen last year, mid last year as well, is that The analysts started to, you know, look into the market to try and understand these technologies, and there was a question around what impact generative AI and generative answering was going to have on search, especially if you remember, you know, chat GPT was its own kind of interface in site that people were gonna go to. There were some predictions of death of search, things like that. But what came out and what I'm sure many are seeing today is that there was a direct kind of quote to say search augments AI rather than the reverse. And what that means is that AI needs search and search relevancy in particular to help process filter and rank the most relevant information that that AI machine is going to act upon. Right? So making good good data go into the system to have good output. So it's no longer garbage in garbage out. I I think one of the interesting points about that is, you know, AI does need surge. And of course, AI grows over time. The more information that is fed into it, the better it does. So I think it's it is better in in December than it was back in June because of the the search input. Exactly. Exactly. And those are there are so many types of AI, which I'll get into some of that. That, is behavioral based, learning based, but not all of it is the same or has the same technique. So it's important to understand what type of AI is being deployed. I did want to just take a minute to talk about search the search bar as we know it because Obviously Coveo is an AI search and generative experience platform. So we're dealing with some of the largest organizations that are, may be more familiar with the value or the power of intelligent search, but not everyone shares this sentiment. I mean, we've done third party tech surveys where, practitioners, you know, tech, tech technical practitioners tell us that they see the value of search, but they really have a hard time getting management to buy into it. I'm assuming again with the whole focus on gen AI and AI, this will change. But I wanted to just take a moment to reframe like wide search, wide search bar. And the fact of the matter is searching so pervasive and so useful in our lives personally and professionally. Just think of all the all the times how much time you spend searching for your research to find someone to connect to learn to understand. And if you're a company and and people are coming to your sites and they're searching, they are looking for something. Right? They are telling you something about their intent. The search bar then becomes your opportunity to having one to one conversation with people at scale and to then be able to guide them, engage with them, make sure they're having a great experience and also use those insights from that user, the things that they're typing, the content they're engaging with, to continuously improve the experience. And so and if you're a commerce or a transactional business, that search bar is even more important to you. Right? Think about searching on popular sites like Amazon or Walmart. Right? We wanna find what we want immediately, and we will transact if it's relevant and meaningful to us. So this search bar really is a value bar that I think really needs to be refrained a little bit for companies. Essentially, the the search bar really is a listing tool for the audience and I think, you know, sometimes the business might think, well, the customers care about this thing, then they realize, oh, we're getting all these searches for this. They care about something else that we didn't realize. Exactly. And, you know, with technology today, we can make that bar quite intelligent and useful for the business. So this takes me to, a little bit more meet and death behind search and search relevance, and specifically the types of capabilities, companies should be looking for an understanding that they need as it relates to relevant search that powers effective gen AI or generative answering to be specific. So I wanted to go through each one of these and if if I need to clarify anything, do please let me know. The first is having a hybrid unified index that is able to ingest and pull information from all of the different systems and application and enterprise has today, and a unified hybrid index is able to do that from a keyword based technique as well as vector. So we talked about semantic search and vectors. So being able to do keyword based search as well as vector based search. In addition to that, you have extensive ranking and filtering options to make sure you have ways of boosting the the most relevant content for end users. And next and very related to that are having the ability to have manual and auto tuning options Things like thesaurus or synonyms, every enterprise is different. Right? Every enterprise has their own product names, their own acronyms. I used to work for big corporation. We had an acronym glossary. Right? And so every enterprise needs the ability to manage and tool results to their domain specific areas. Of course, we need to talk about security when it comes to AI, genai, or enterprises in general. And if you think about it, again, enterprises thousands of applications, each of those applications has their own type of permission access controls. So how are you able to standardize and ensure only those who have access to your thousands of systems get the results that they should be able to see and and nothing more. Right? When you're interesting on that and that you're saying possibly that the search box would bring up different results if you're a customer or an employee or administrator, and they would see different results. Absolutely. And they should see different results. In addition to that, we, you know, you have to think about the connectivity and silo nature of enterprises. So having a system that has out of the box connectors and native integrate integrations to popular enterprise systems like Salesforce, Adobe, Sitecore, Right? That helps you to be able to establish interoperability and to also augment those systems again with Greek unified search, AI powered search. If I move to the right hand side of the column, we then get into content freshness, right, which was a big deal when it came to AI and Gen AI and It processing the most relevant information and resources. Think about how much how many new content you know, knowledge, documents, and data enterprises are creating on a daily basis. So systems need to be able to refresh, you know, quickly to stay relevant to the business. This next one, I feel like, is super critical and strategic, and it's around giving the business control over the results. So, you know, let's say you're again, your your product business, and you want to feature a certain high margin product, or maybe you wanna feature an item where you have too much suck and you want people to buy it. Giving businesses the ability to manage or override AI and the results is absolutely critical to helping that business achieve their strategic business objectives. And we talked a little bit about, the behavioral side James around, AI that's able to learn and and understand and and track success on thousands and millions of data points. That's where behavioral AI models come in to positively reinforce the overall system, the overall and generative AI system. Right? And we didn't talk about it, but personalization or that ability to to use a person's real time behavior, the things that they're clicking on in real time, plus their historical information, maybe it's past purchase, purchase orders or past, activities, the ability to use both of those to really offer real time in session personalized rec recommendations, whether that's content or products. Again, brings a level of intelligence to help make that person success soul and hopefully transact with you. Last but not least, it's the need for analytics and insights into your overall system performance. How do you know search is working? How do you know your AI models are working effectively? Being able to draw all of these insights and data together in a system in a closed loop manner to continuously improve really is what's going to make that difference in building a successful generative AI solution. I I think that last piece, the analytics and insights, it is a is important as any on that list and that companies do need to be looking at what the search results are and what what it says about what the customers want. Exactly. Beyond this, and that was a long list of features and capabilities. So thank you for enduring that with me. Beyond this, I also wanted to point out few other differentiators. For those of you, really try to assess, okay, why AI search, or what are, you know, differentiators, and that set a good or great platform across from an okay platform. And what I would say to that are the are the items that you see here. A great AI search platform is really gonna do the heavy lifting for you. Here it says managed AI solution, but what this means is you have the power of many different AI models that are in and out of the box approach. So you don't have to hire an army of data scientists which many companies, you know, can't really afford to do. You want a system that is closed loop self learning reinforcing, we kinda discussed that in the last point, but of having a closed loop system from front end to back end. Continuously improving is absolutely critical. We know enterprises are large, many departments, many different stakeholders. So having a system that lets you take a multi use case approach for improving your website, maybe it's your intranet, maybe it's the service stuff, or sales team, having that flexibility to do that is important. Also, truly being agnostic when it comes to your user interfaces, the content you're processing, the applications you're working with, oftentimes, the big, you know, platforms that get invested into, cannot work with content outside of their ecosystem. And so this is where, you know, an AI search platform comes in to take that agnostic approach. Should a matter where where the content is, where what system it resides in. I talked about business controls on the last side. I wanna emphasize this again. A system, you know, is is effective and powerful with all those capabilities I mentioned. In addition to giving the business a troll to promote override and, you know, achieve their business objectives, especially if you're a product company. It goes without saying the importance of enterprise scalability. You want a system that is enterprise grade can grow with you. You don't have to worry about it's robust. You know, basically built for the enterprise. And and also last but not least here is best of breed technology, which we hear from companies all the They wanna work with and buy from vendors that are gonna be constantly innovating, staying up to par or ahead of standards, really trying to avoid that technical debt or tech lock in. So hopefully this gives you an idea of other different changers in addition to the features and capabilities I talked to on the last slide. I wanted to change gears here and actually talk about generative AI or generative answering in action. And what you can see here is actually one of Coveo own customers. Zero, who is a worldwide accountin accountancy software company, Zeros been a Coveo customer for the past eight years or so, and they initially came to us help improve their online, online support community. Here you can see it's Zero central. So they use our AI to really help users find what they need, help them self-service, help them self resolve, and they expanded over time even to use Cobail in the agent experience side. So if you think of yourself as a customer going online to get support on a product, Ideally, you wanna be able to self resolve and not have to call the help desk, but should you have to speak to the help desk? You wanna make sure they know who you are, They have some idea of your online experience, the things you just went through to try and self resolve, and you want them to be empowered to be able to help you solve that issue rather quickly. So that was our history and our story. Was zero over the last eight years or or so. Mid last year, we did release our generative answering capability, and Zero was one of a few, of our customers that decided to beta test try generative answering on their own sites. And so you can see here they actually try generative answering on that same community portal. So you can see someone is asking how do I update my subscription payment details and generative answering is giving them a question what steps with, the the citations or sources of that information and allowing them to give the system feedback whether that answer was helpful for all great important capabilities that, again, are helping to accelerate a customer's time on the site and their ability to be successful. You see some numbers on here. Those were the most impressive for us, which is By implementing the solution, Zero was able to increase the self-service re resolution rate by twenty percent. Meaning, making customers twenty percent more successful or saving their time so that they find the answer quickly and can move on to the task at hand. In addition to that, the ROI they were able to experience was significant within six week they were able to see a clear ROI from this generative answering solution, as well as reduce the average time to resolution for any calls headed to the the service desk or help desk. So a lot of great success here, and this is on top of a solution that already was leveraging a few AI models. So this is now generative answering, providing the leap more in productivity and and valuable at a time to end customers. What did a a question on that? I'm I'm curious. So This is JNAI answering. Does that mean that it would go hand in hand with, like, straight retrieval kind of answering or if you're if you're using JNAI answering That's what you're using. How how do they combine or or not? It's a great question. We really see generative answering as Another model. For us, it's actually our eleventh AI model. But to your point, there are other types of AI models that would give you an answer, a direct one to one answer. What we could call is non generative answer. We call that model smart snippets. I know others call it other things, but there are multiple methods or techniques that you can put into place, and generative answering is just one of those. So should zero say, you know what? Don't wanna generate an answer. I want that direct answer in that knowledge article that is ten pages long. There's models that can bring that one to one, answer for an end user, and we and and we offer that as well. Right. So hopefully you can see the power of generative answering when done right, and all of this would not be possible without all of those search foundations and capabilities that I mentioned, all of those search features and capabilities are necessary and generative answering is built on top of that rich, robust set of features and capabilities. And one done right who get great results. The other thing I just wanted to highlight here was, you know, what we looked at was a generative answering inaction on Darryl's support website. But you can imagine, generative answer can be on any full search page experience. It can also be within your service management application or your CRM search, and it can also exist within a bot like experience, we have our own in app product or in product experience is what we call it. And you can see generative answering in that real estate as well. So I'm sure this to say, I mentioned earlier, you know, being agnostic and having multiple use cases, multiple UIs is really important when you think about the fact that enterprises are catering to a variety of different stakeholders, so that flexibility really does matter there. When I just talked about this a little bit, James, but I thought it deserved to be, highlighted a little bit more. And that's the fact that generative answering really is what I'll say is one of several search techniques. And what I mean by that is if we look at the image here on the screen, generative answering is one, and there's two others that exist. Lexical search and semantic search. Now, lexical search is just our basic keyword search, which many of us still use today, we don't always need a generative answer to get to, you know, what we're looking for. Sometimes we do just want that document. Sometimes we want that link. Right? The generative answering is not always needed. And then with semantic search, it's, you know, going beyond keyword search and being able to understand what we meant, when we only provide limited information into the search bar. But this whole category on the left hand side, to me, is a bit reactive, meaning it requires a user to search. Right? You have to search to find an answer. But we have models today that are more proactive and don't require a user to search. And those are recommendations models. Think about content recommendations, product recommendations, people like you also viewed, people in your function also read. These exist as well. And so my point here is there are multiple techniques that companies need to consider for their use case for their enterprise. And while generative entry, of course, is one of the most exciting. It's not the only one that can be used and it's not the only one that people want to use. Now this gets me to, a bit of what I was saying earlier, James, and that's that the reality today is that enterprises are a bit overloaded with a bunch of silo platforms. And for those of you joining us, I'm sure you recognize these acronyms, everything from your or ERP to your content management system as your marketing automation platforms. Right? Each of these platforms are absolutely critical and important for that department, so very core systems and applications. But the reality is that your end user searching, whether they're an employee or a customer, they don't really care about those systems. They don't really care about where content lives. Right? They wanna find the information they're looking for as fast as possible so they can get on with the task at hand. Furthermore, your back end team, your developers, your IT, your systems managers. Okay. They care a little bit more about these systems because they're managing and maintaining them, but they obviously wanna be working with the best technology out there. They want innovative solutions. They wanna figure out how to make these different forms interoperable. And and to that, that's where I say an AI search type platform really can bring a lot of value. On the front end, it's helping delivery connected, intuitive, intelligent experience for end users. It's helping them find things, the cross sites and systems. Again, for an end user, They don't need to worry about the detail. They just wanna find what they need, find and discover. But what's in you know, you could amplify and augment the systems you have in place today. The reality is the systems you see here You know, they've been there for a while. There's millions of dollars being invested into them. They're not gonna go anywhere, but just because they're there doesn't mean that you can't breathe a new life into them in brain capabilities like unified search, AI recommendations, generative answering, all the things you covered before, These can be used to amplify your existing systems, but also to unlock the value that's stuck in them, right, to to unlock the knowledge and information and make it accessible to others. And lastly, for your dev teams, you really wanna future proof their experience by future proofing you know, your your future proofing your your tech stack, you're helping your dev team upskill a little bit, you're amplifying your systems, you're allowing them to work with best of breed technology that in a way up skills them overall as well. And so these are the points that I think are in favor of introducing an AI search platform that really is going to connect, really going to amplify all these different platforms and improve the experience to your front end users as well as your back end employees. So, really, overall win win situation for all for all sides. So, Bonnie, are you saying that, for an AI solution to be effective? It should need to be hooked into all those silos like ERP or CRM or or not necessarily? Not necessarily. I think my point here is that, oftentimes companies think of these stems in silos and only purchase or purchase in siloed platform mentality, but I was really thinking cross platform, which really thinking cross enterprise to to bring the value on the richness there. Mhmm. Okay. And with that, I did wanna just do a little overview of Coveo, to help people understand, the robustness of capabilities that go into an AI search and generative experience platform. And so what you can see is kinda working our way up the connectivity or all those applications and systems in the enterprise. Right? There needs to be a way of connecting them and you finding the access to that information securely so people can find what they need and really have that single source of truth across the enterprise. There's many ways in doing that. We then get into our, our AI models, which, you know, we talked about generative answering. We kinda mentioned a few others, but today, we have over ten models actually eleven to be exact distinct AI models that are in use by customers throughout the world today, and those models behavioral learning from users and making recommendations. They're personalization based. There are some that are LLMs that are non generative, which I spoke about smart snippets as well as generative with generative answering. And then there's our front inside, right, How are you able to improve those experiences on your sites and within your applications? And that's what Coveo is doing every day for our customers really bringing a robust portfolio of solutions and capabilities to improve experiences on your dot com to improve your commerce portal, service portals, intranets, and agent experience. So this just gives you an idea of, you know, again, there are a best solution capabilities that go into a platform like this, and the goal again is to help manage those information challenges and make that genai effective. Again, without great search, you're not gonna have great generative experiences. So wanted to look at it at the chart. It looks like I see the large language models down there. Is it correct if I figured if a if a company's gonna purchase Coveo, they, of course, are gonna customize one of those large language models? Or how would that work? Yeah. It's a great question. There's many ways that different companies do this. I would say Coveo was a more of a managed AI platform, which means we have a team of over two hundred r and d data scientists that are honing our models for a horizontal use. So when you buy a covalent platform, you get these eleven AI models out of the box. And for you to set it up, it's more of a few clicks and drags and drops than coding So we do all of that for our customers. The large language models we use today, we we do use OpenAI, We have a way to, like, connect to them, but, it's all secure within our platform, which is bought by our customers on a on a cloud cloud native basis. So you we're not taking, like, weeks of of customizing a larger language model. Oh, no. Absolutely not. We've we've sort of refined our models to work with the, you know, enterprise knowledge and information. And so a customer can get started rather quickly with our models. It could take them even a couple of days. All it all the main steps, I think, would be connecting, your data sources. So knowing what systems you're really trying to Unify access to knowing what users you really wanna improve that experience for. Is it your employees? Is it your customers on the website? And then, you know, applying the models to to your different search experiences. That's just as easy as that. So we can get complicated. We can get basic I think we really tried to give options for helping people get started, luckily. Good to know. Yeah. And on that note, I think another, you know, important question that comes up quite a lot is the build versus by debate. This is in relation to, you know, search platforms. It's definitely a hot topic for AI and GenI in general. I think everyone is figuring out the right approach for them based off of their business, but I wanted to just offer a few considerations On the build side, you know, I think it makes sense to build your own l l m's or generative solution. If you have a large team of AI experts, right, and some of the largest, you know, tech giants out there do. And so, they're gonna be building this, themselves to incorporate into their solution portfolios. Also, if you're building AI or GenI as part of your core competency, let's say you have a custom financial application that is core to your business and how it operates, you might wanna, you know, build on your account. It can be a little less dependent on on someone else or something that's very core to your business. If you're also just looking at a point solution like one off need, probably building on your own something simple is is okay. And if you're not thinking, you know, enterprise wide or don't have scalability concerns, Gold approach is probably good for you. Reason to buy or consider an AI search platform, first for most would be, of course, if you have limited resources and especially limited AI, expertise resources, which you know, not not a lot of companies can can spin up. Right? It takes tie, dummy and expertise, a lot of investment to do that. So partnering with someone or working with a vendor is a is a good approach. I would also add in, if you wanna avoid tech lock in or technical debt. So a lot of companies today, right, they buy into an application, And depending on the vendor, whether they're enterprise grade, you know, public, their road map, chances are they get stuck with technology that's, you know, eventually becomes outdated, isn't as modern, and that does slow companies down. So if you wanna avoid tech lock in and and go to a more agnostic, right, best of breed provider, which we see here on the list as well. That's that's why you would go. You know, third party. And also, last one I'll cover here is if you want an enterprise wide solution, right, again, I mentioned naturally enterprise behaviors are buying more siloed, more departmental. And that can be you can argue that has caused these data challenges to some extent. But if we wanna start to resolve those data, you know, challenges those data silos and thinking enterprise wide and the ability to, you know, make something usable cross departmentally or horizontally is absolutely a good cause and a good reason to buy a solution rather than build it on your own. Yeah. If I could speak for most companies, I think they're the of, oh, we'd we'd much rather buy than Bill because Bill creates his own headaches. But, yeah, who knows every company out there? Yeah. Fair. And, you know, there's a lot of there's a lot of debate on this one. But, yeah, you know, I think it's different for every business. So everyone has to decide what makes the most sense for them. So that we're getting near the end of our planned presentation I did wanna share just a couple of resources if you're interested in learning more. The first is the blog, that I actually authored called genAI is only as good as your search. It goes up to a little bit deeper. The it goes in a little bit deeper on the topics we discuss today. So if you wanna give that a read, we'll drop the link, here. And then we have a, what I'll call a spring innovation release On March twenty seventh, we call it relevance three sixty. But this is where we showcase our latest innovations, where we are investing in. And we also hear from industry leaders. Excitingly, this time we'll have, guests from Forrester and OpenAI to talk about search. Generative answering and the importance of relevance. So I I'll I'll drop the link and, would welcome you all to to join us. With that, James, wanted to thank you and thank those of you joining us today. I will now pass it over to you for any questions. Yeah. A lot of good stuff on it. Absolutely. I have just a few questions, and I do invite the audience, please, do put two input your questions and we'll we'll do our best to get back to them as soon as possible. Even after the day of the recording, no, please, input your questions. Well, I was wondering what I need to alright. So say a company purchases Coveo, is their data secure? Because you certainly, there's a certain data merging. Is there not so I'm concerned? What about the issue of data privacy? That's a really great question. What I would say to that is we are a cloud native, solution. And, what's interesting about our platform is there's no data migration. Right? So your data will always live within your systems, your core systems of engagement or record. But we are indexing and unifying the data that you you tell us to to process. And it's done through a a private secure cloud that our customers own and manage, independently. Right? So they're effectively buying us as a SAS based solution to rate. So it's absolutely secure. We uphold the the highest security standards. We even have a HIPAA cloud. So when it comes to data privacy, we really take that seriously. And one thing I think I would like to call out is we also do early binding security, which means during this whole indexing process of when we are, you know, processing and ingesting data into our into our cloud, we take permissions from those source systems, again, ensuring that people never get access to things are not meant to. So, we definitely take security seriously. An important issue to be sure. Alright. What about setup? I mean, also concerned, is this could be weeks to set this up with the whole team? Or what what what does it mean to set up the COBRA solution? Yeah. You know, it's a great question. I think it really depends what your use cases. For example, you know, I'm a big fan of our digital workplace solution and really helping improve the spirits on intranets, right, and findability and helping employees find, the right information to do their jobs. But, you know, the intranet is one of those, applications that is kind of owned and managed by multiple departments. So it gets complicated in terms of getting people to buy in, You have a variety of content on there. Even budgeting and decisioning gets very complicated. So I think, you know, that's a more lengthy process only because people our customers have to work a lot more internally with their different stakeholders. But when we look at a use case, let's say, for a website, something as simple as, improving, you know, findability or discoverability on your dot com, your website. While marketing typically owns the site, they own the budget. Right? And they can they know their content mostly. There's fewer internal stakeholders to have to win over And so getting started again is as simple as knowing who your audience is that's going to be searching. What's the content they'll be searching for? And then spinning that up within the Coveo platform, which, we are constantly innovating to make it easy to use to make business users self sufficient. So there's less reliance on IT. So all that's to say is it's rather quick. I think it's more of the internal stakeholder, management depending on the use case that that makes things longer and more complicated. Well, I guess, on the other same, no, mean, what if the platform's been up and running for o six months or so? And suddenly the company has a whole new product line. They wanna make sure that's part of the search results. Is that a few clicks or how how would I guess, really, I'm asking about maintaining it. What what what is the issue with maintaining it? Yeah. I would say it's a few clicks. To add a new data source or to ingest a new set of data. It's pretty quick. And then our our users are typically maintaining it. The best ones are, you know, maintaining and learning and looking at those analytics, right, on a regular basis to make sure that People are finding what they need and what let's say there is a content gap. Let's say there's a no result search. Well, that's indication that maybe you're not getting in just that data, which means you're not amplifying it with AI. So I think having those controls and insights help to really improve the experience. Okay. What about issue of running and you talked about this a bit, but I just I'm I'm not totally clear on this. The issue of running generative AI with the straight retrieval with the other search ways, search technologies. Can a company tune that? Does it say we're all geni gen AI where we're gen AI mostly, but we do some straight retrieval, or how does that work to combine all that, or or is it all generic? It's a really good question. I would say today, we have a couple of options. One is we actually offer a toggle box to turn it off from an end user perspective. So let's say you're searching on a site and when you're searching, generative answering is working in processing. Right? And let's say for you, you're like, I don't want generative answering, you know, I want to just find the link or the document directly, as an end user, you can toggle it off so that you don't have to deal with generative answering. And over time, because Coveo does offer you know, keyword based search, content recommendations. Over time, the systems will get smarter, so they'll be able to tell, like, a two keyword search into the bar is probably not a generative answering type question. That's a basic direct retrieval question. So we don't need generative answering for that. But if it's a question that's, you know, a longer form, something a little more complicated that involves comparisons or steps or set up, then the system will know, okay, that's the candidate for generative answering. From a business perspective, it's always up to them on when they're using generative answering, and they shouldn't really think about it because it's not cheap. Right? There's a premium for generative answering, so you wanna make sure you're using it in the places that are bringing value to your end users or your customers. Interesting. Okay. This is really, really fascinating. Juanita, I think you said it. I learned a ton, and I I wanna thank our listeners a day, and I also don't want to remind them that we welcome your questions. Even if the the, day of the recording, you know, we'll try to get back to you as soon as we can. So please put your questions at any time. We need it. Thanks again. Thank you.