Name: Is Your Business Ready for AI-Powered Copilot Solutions?
Uploaded: 2024-11-05T17:06:01.600Z
Duration: 48 min 4 s
Description: Is Your Business Ready for AI-Powered Copilot Solutions?

Transcript for "Is Your Business Ready for AI-Powered Copilot Solutions?": Welcome, everyone, to today's webinar. Is your business ready for AI powered Copilot solutions? So I'm Rav Simak, strategic partnerships manager at Nasuni, and I'm really excited to be moderating this session, with 2 great speakers. We've got Jim Liddle, chief innovation officer of data intelligence and AI at Nasuni, and then we have Mario Henriques, manager of Microsoft Integrations at Seagal. Seagal are a Nasuni partner specializing in the energy sector and really thrilled that they could be with us once again today, first of, second of many events that we've we've done, this year. So we have about 1 hour allocated for today's webinar. And on the agenda today, Jim's gonna kick things off with a a brief overview of the Nasuni file data platform. He'll move on to how Nasuni integrates, you know, with Microsoft Copilot, then next, Mario is gonna be challenging you and and you know whether your organization is really ready for CoPilot before going into a customer story and to wrap up you know we'll use the last final 15, 20 minutes for an interactive Q and A. Just a few housekeeping items, so you can use the q and a tab to ask any questions that you have throughout the webinar or you can save them up for for the interactive q and a. Totally up to you how you decide to do that but we'll be monitoring the questions throughout the session today and we will be running a poll as well so please keep an eye out for that. I'd just like to encourage everyone to participate. And finally there's a downloadable document which Jim will be covering as part of the the session, and you can download that just by pressing on the docs tab. Alright? So that's all the the housekeeping items that I have. I would like to hand over to our first speaker. Jim, over to you. Thank you, Ralph. I'm gonna be talking today about, what Nasuni does with, Microsoft Copilot. I'll be looking at both Copilot Studio and or 365 Copilot. I'll be talking about what we do in terms of the integrations between the two. And then I'll be handing across to Mario who will talk about some real world scenarios in terms of what we do, inside of companies, you know, with Copilot and how complex some of those integrations can be. First, let's do a quick Nasuni recap in terms of, you know, what Nasuni provides today. So for those of you who don't know anything about Nasuni, we provide, a hybrid cloud file infrastructure that's distributed either regionally or globally or both. What that means is is that we have a a global file namespace where files are stored inside of the object storage provider of your choice on cloud, whether that's kind of Microsoft, AWS, or Google. And the edge devices, you know, which can be, local to your office, they allow you to get data where the application's needed. So that solves the real data locality or data gravity problem in terms of a lot of apps don't or can't wait for data to be downloaded from the cloud. That all presents, a file, an SMB or NFS interface. And because we often have companies who are working on similar datasets between offices, that's not unusual in a lot of sectors, the AEC, energy, you know, media marketing, then we handle all of that global file locking, you know, between the files that are being worked on, all of those distributed file locks, and we handle that transparently. And in the background, you get a continued version stream. So we're automating all of the snapshots, you know, back into that global file namespace. Why is that important? It's important because in, in the, in the case of you getting some form of threat actor, or ransomware problem, then you'll want to not only mitigate the blast radius, which we do with our ransomware protection, you'll also want to be able to restore those files as quickly as possible. Now, you know, one of the great things about Nasuni is because of the way that we do those version file streams, it's all about metadata and pointers. And that means that we can restore millions of files in seconds, not weeks or months. And ultimately, because we have that global file namespace, when you're pointing towards AI AI services, all of that fresh data that's coming from those edge devices is actually going back into the global file namespace and then is available, you know, almost immediately to those AI services. So the architecture is distributed but and federated, but it allows you to get that data back to those AI services in real time, which is actually quite a difficult thing to do. 1st, I'd like to talk about what we do with Copilot Studio. So Copilot Studio is essentially what Microsoft provides to allow companies to build custom Copilots. That's Copilots that they go away and build and manage themselves. Copilot Studio is really, what you see is what you get type of web based development studio. And it's integrated with something called Power Automate, and which makes it very, very powerful. You can add, you know, Nasuni datasets there either, you know, just by dropping datasets into it where, those datasets get indexed and vectorized, or you can use a graph connector. You know, I won't go too much into that right now because we'll be talking about the graph connector when we talk about o 365 and Copilot. One of the great things about Copilot Studio is that it's very good in from institutional knowledge perspective for very discreet use cases. So where I've seen customers use it, it's to do things like, allow the hitch the HR department in the HR teams channel to be able to chat with the AI solution around perhaps customer contracts, or for the legal department to do something similar. I often see it on project based initiatives where people wanna put legislative documents or project documents, make them available to the copilot and the teams channel, and then the internal people working on the project and the customers invited to the project, both can then interact with AI on that project documentation. It is very good for those types of use cases. And, of course, it's also really good for, I would say, the the the low hanging fruit use case, which is the kind of sales and support chatbot. Since we, first announced the, the ability to work with Copilot Studio, a lot of things have changed. You know, there's a lot of new things inside of Copilot Studio, and we'll be talking about some of them when we do a comparison between, you know, what is all 365 Copilot and Copilot Studio. But one of the things that's quite new now is Copilot agents. It's currently available for you to go and nominate your company to get access to it in preview. And I thought it would be at least worth spending a slide on understanding what is the difference between agents, plugins, and actions. And the really the main difference between Copilot Studio agents, you know, you know, what's in preview right now, and Copilot Studio actions and plugins really lies in their scope and their functionality. Copilot Studio agents are really stand alone AI chatbots. What does that mean? It means that they're very task oriented, they can be customized for very specific purposes, and they can be deployed independently to do that task. They can use their own foundational models, and their own orchestration, and they're very separate from the kind of Copilot Studio and 0 3 65 Copilot. On the other hand, the actions on plug plugins are really components to extend the capability of your existing copilots that you created. They add specific skills or they add functionalities to enhance those copilot capabilities. And what we're gonna start to see is a lot of those agent type copilots that, are able to do those autonomous tasks. Perhaps even in copilot flows, you know, where, a user may ask a specific question, and, actually, that might get handed off to one of these agents to actually go away and fetch the answer. Next, I'd like to talk a little bit about, Copilot for, 0 365. This is something we've only just, you know, recently announced. And that's different to Copilot Studio because Copilot for 0 365 is essentially what Microsoft provides if you buy the Copilot for 0 365 add on. And it's a combination of, you know, the large language models in the background, the Microsoft Graph, which is basically taken in your data and vectorizing it and indexing it, and the Microsoft 365 apps, which act as the client to allow you to interact with the dataset. The O365 Graph Connector integration is really made possible through something called the file share graph connector. And the file share graph connector is really the glue that allows the Nasuni data to be continually indexed and available to 0 365. And that works a little like this. First of all, you've got obviously your Nasuni data, and then that data is made available through the file share graph connector. That gets, indexed through Microsoft Graph into the semantic index. It's available alongside other or 365 data, and ultimately, that becomes available to all of those apps in that ecosystem. That means that when you, are really working with your dataset through or 365, you are working not just with or 365 data, you're working with your Nasuni data. And really adding that external Nasuni data to or 365, it expands the knowledge base that's available to or 365. So it's really providing a much more comprehensive view of your organizational information, and that ultimately allows, Copilot to access and reason over a broader range of content. Why is that important? Because you'll get back better informed and more relevant responses. The other thing to to point out is it's not just about 0 365 Copilot, the add on. When you index the data through a graph connector in this way, you also get it made available to you through Microsoft Search. Microsoft search is available in the 0 365 apps, but it's also available on the Windows desktop. So when you, when you're indexing, you might not get semantic search if you're not using, 0 365 Copilot add on, but you will get keyword type search through Microsoft search. And we actually have seen, some of the companies that we work with average into the use of something like 0 365 Copilot by indexing their data first, you know, sort of available through graph in the semantic index and available to, to Microsoft Search. And like I said earlier, it's really about improving the accuracy and relevance of the dataset. The ability to have more of your, organizational information available to the large language model, so that it can come back and give you more relevant answers. Now one of the questions I get asked a lot is, what about, security? I really don't wanna be putting some of my datasets into something like 0 365 Copilot through Nasuni if somebody asks a question and they get access to HR information, for example, or they get access to legal information. But one of the nice things about 0 365 Copilot, is that you can actually, map your active directory, with it when you're doing the setup. What what does that mean? It means actually that it will start to utilize the access control lists and the groups that are inherent within the active directory on the files. So that when a user actually asks question, it's only based on information that they have access to. That's honored through, 0 365 data using Entourage ID, and then it's honored through the files that you have, you know, with, the SUNY through the fact that when you set it up, you bound it into the active directory. And then it the graph connector really graphs a combination between the 2 so that when you ask the question, it's able to filter back and make sure that the user doesn't get an answer that they shouldn't get because they don't really have access to that information. So what are the difference between integrations today? Custom copilot, it's really something that the organizational builds and manage. So you gotta remember that you're responsible for it. You're responsible in updating the information for it if you're not using the Graph connector. You're responsible in terms of tuning it as well, because there is some tuning that perhaps needs to be done when you decide to build these custom copilots. All 365 Copilot, you get out of the box. So if you've got the all 365 add on, through your all 365 subscription, then you can leverage all 365 Copilot. You can change the interface. You can change the icon. You get what you get from Microsoft. And I see actually sometimes companies using both. Sometimes it's not a case of 1 or the other. But why did companies gravitate sometimes towards 0 365 Copilot? Sometimes because it's frictionless, and it provides a single pane of glass into some of that enterprise knowledge, and you don't have to manage the clients. And, obviously, end users get used to that interface as it gets rolled out. So it's it becomes much more frictionless from an IT perspective to be able to manage it. And actually, many of those teams spend a lot of their work in life inside of, teams or inside of all 365 apps. So having the search ability and the ability to q and a with the documents to hand all of the time, that's another reason why, you know, companies tend to gravitate a little towards that. The the other big thing around this is, file sizes. So if I've done this webinar maybe even a month ago, the file size the file size for custom Copilot would be 3 megabyte because that's what it was when it came through the preview. It's now being raised up to 512 megabyte. So what you get in terms of the data you're working with is really what I would call knowledge documents. That's textual type documents, PDFs, text files, XML files, Word documents, Excel documents, PowerPoint, but not images today. And in terms of what you're doing when you're working with Microsoft 365 Copilot, exactly the same knowledge based files. But right now today, that's limited limited to a 100 megabyte file size. So that's something to be aware of. I actually think over time, we'll see that file size raised on the o 365 Copilot, perhaps bringing it more into line with the custom Copilot. But today, those are what it that that's what it is. And what are the top Copilot use cases? I can only speak to what I see a lot of the customers doing across industries. I think the the biggest one is what I would call institutional knowledge. It's just the ability to be able to q and a with certain document sets or across document sets. That is still, a big deal. I mean, if you think about the amount of unstructured data that gets accumulated by an organization, it's very large. You know? And Gartner talks about it being dark data. Data where actually you only ever see the tip of the iceberg of that data because the rest of it kind of vanishes, you know, inside of your storage. So the ability to be able to q and a with some of those datasets is a big deal from an organization point of view, and it can it can actually tackle many use cases. I can give you one example, that I was working on recently in the AEC space, you know, where they wanted to be able to pull their, tender proposals, into 0365 Copilot so that when people were using Microsoft Word and they were doing a new tender, they could leverage what they had already done from previous tenders. So they wanted to index their circa 20,000 tenders to, you know, make it available so that they could ask questions of the Copilot to help them with a new tender. And those types of discrete use cases are great use cases for the use of things like Copilot. And of course, again, I still see companies replacing what I would call sales and support one dot org chatbots, the ones that were only built using static rules. A lot of companies are still taking that low hanging fruit and replacing them with something much more intelligent. What I would say is that, use cases are very sector specific. So if you looked in in the energy sector, for example, you'd end up with some very different use cases of Hamco pilot, you know, than you would do across some of the other industries. I won't go into these. You know, I'll pull some of these out. What I would say is that, you know, when Mario talks about some of the use cases they're looking at, you'll get a good insight into what, you know, what some of those use cases can be and what how complex they can actually be, how complex you can get with some of this technology. Points to be aware of. In terms of the, Nasuni integrations, we don't charge anything today for the use of the Graph connector. Graph connector is provided by Microsoft. We do provide a professional services capability, which can be a paid capability if you want us to take advantage of providing, you know, the setup for that integration. But actually, we provide you a lot of documentation. It it's fairly straightforward to do and to set up, and particularly on the 0 365 side, and it's definitely within the capabilities of most companies to be able to do that themselves. The other thing, that's worth mentioning is, again, if I had done this webinar even a month ago, Microsoft would be charging you $1,000 per 1,000,000 files in the semantic index for 0 365 because that's what the price was. Today, they've lifted that, that price. So now you can store up to 50,000,000 files in a tenant for 3 for free in the semantic index. That means you can index up to 50,000,000 files from your Nasuni into the semantic index into o 365 and Microsoft. There's no fee from Microsoft from that. One thing I would say is that the limit for a tenant today is 50,000,000. So you're actually at the the cusp of that 50,000,000. 50,000,000 might not seem like a lot. When you're talking about the your file sizes, you might be thinking, oh, well, I've got a 1,000,000,000 files set about on minus 0.9. But the reality is, I would always advise people to be aware of the garbage in, garbage out principle, which is a lot of those files that you have, you probably don't want to index and make available inside of these copilots because they're either old, out of date, contain versioned information, contain information that's not relevant, or contain confusing information that actually confuses the copilot. So I would always advise, all of the customers who who are gonna build these copilots out to go through kind of a relevance check, to do an ACL check just because, there are access control lists available on the files doesn't mean they're the right ones, and sometimes that can surface things very quickly when people ask questions. You also wanna be able to go in, you know, just look at the format consistency of some of your files. You wanna go through a process, you know, a consistency process and a curation process just before you put these files in. Now some industries already have this in place, and they're very good at doing this. You know, AEC actually comes to mind with that. They've got very fixed directory structures. They've got gold master copies of data. So I would always say, look at this, you know, very precisely, but you can kinda see why actually getting to that 50,000,000 file limit might not be as easy as you first thought. In terms of the, the resources that we have available, we have, initial solution overviews. We've got a very technically deep white paper on how to set up, the 0 365 Copilot and the and the Copilot and Copilot Studio. We've got demo videos on each, which have just been brought up to date from the Copilot Studio side so that, because all of that interface changed, and that step you through how to build out an institutional knowledge, chatbot, you know, with your data. We also have the professional services resources I talked about earlier, and we make all of this content available either on our website or something we call the SUNY Labs on our GitHub. That's it for me. Appreciate you all listening. Really excited now to hear from Mario, and what they're doing at Sigal. Mario, over to you. Thanks, Xin. It's, super interesting to see and to confirm that everything that Nasuni has been building on the infrastructure side on the past years, it now makes so much sense and it's so much relevant for this new Copilot's world. So, yeah. So my my presentation, I'm going to challenge you a bit. And the first challenge, that I have is, are you really prepared for what's coming for this new, Copilot, new era? Well, first of all, just a quick introduction myself. My name is Mario and I work for a company as Jim mentioned called Sigal, and and Sigal is a company, that that focus, primarily on the energy sector worldwide. Some of the numbers, from Sigal is, you can see it on the screen. There's, we have customers, basically, worldwide. We are spread at all over the place, in all all the regions that are relevant for us. Working, today with more than 700 different energy customers. Their quarters is in, in Oslo, Norway, but we are, present in, in a lot of different locations. Why is Sigal, different? It's because we do have a perfect combination of, skills, services, and expertise on the energy. And I think that's, unique, worldwide, and that's why a lot of these customers rely on Sigal, to, to collaborate and to, achieve, the next wave of, innovation. As I mentioned, we are, globally available, and we do work with, the biggest energy customers worldwide. And and as you can see in the map, it's pretty much, all over the place. But let's go for the interesting thing, Copilots. So Jim briefly mentioned, Copilots and how Nasuni is tackling this, this new, evolution in the technology. So, of course, we know that Copilot is not but, it it's not new. AI is not new. What is new is actually the the technology that allow humans to interact with AI, and that's the large language models. That's the new technology because, the the AI technology per se is, quite old in a way. However, it's it's, interesting to mention that what it makes Copilot possible is the large amount of data that Nasuni, is, is, supporting, which is now exponentially increasing. So Copilot is not only using a lot of data, but on the other hand, it will generate more digital data. So very soon, we will face, overload of digital information, and that's where, things will start to be fun. But, yeah, copilots. There are so many different flavors these days. We know that, all the big companies have invested, a lot of resources and money to to bring their own copilots to, to the market. There's there's a lot of different different aspects on each one of them. I will focus on the Microsoft Copilot because it's just to follow-up what, what she mentioned in the presentation. But let me just, first, say something about, what I really think Copilot is. Since I was a kid, I always admire Ironman. Not because he was a talented born superhero with natural super skills, but because he managed to enhance, in a way, his capabilities by using technology. He created his own superpowers, and, that's exactly what I think we are now creating with these new Copilot, or Copilot experiences. We are, giving users, their own superpowers. So, of course, there's always the concern about, our copilots going to, steal the workplace for a lot of people. I don't think so. I think what is going to happen is that people will do, will will go to a different stage, which is these new superpowers will allow them to to do way more and more efficient way, giving them way more time to do what they are good at and leave all the recur recursive tasks and, a little bit more, time consuming tasks, and Copilot will be able to optimize system on that. And by the way, this is our men number 1 magazine or comic, released in 1968. So, just out of curiosity. As I said, Microsoft is 6th Copilot. So, Jim briefly mentioned mentioned it. And most of you have, probably seen this slide from Satya when Microsoft launched Copilot. How does Microsoft Copilot work? It's basically, as I said, the enabler is the large language model technology. Copilot is available in all Microsoft 365 applications. And the key technology, the the the core technology behind the scenes, which allow us to actually, get the right data at the right time is Microsoft Graph. But there there is a small a very small detail that, most of the people, might not be aware, which is, as you see on the bottom left corner on this slide, is that Microsoft Graph can reach the content and the data that is available on your own Microsoft 365 tenants. What does this mean? It means that your superpowers are somehow limited to the content that exists in your own company. And this is where Nasuni and then Nasuni, infrastructure becomes super relevant. We want people to have the best superpowers in the market. And, of course, we know that most of the customers don't have all the data they need. The data that allow them to make the right decisions at the right time inside their own tenants. So they need to to to increase the scope or they need to increase the reach of their Microsoft Graph by bringing data from other different platforms and data sources that are, relevant for them. But I'll come back to this in a second. So when when we, when we talk with our customers about Microsoft 365 Copilot, this is, interesting workflow that we get from our customers. There's a lot about security. There's a lot about, privacy, data, information protection, sensitivity labels. There's a lot of, details that most of the customers mentioned, but they typically forget one that is the most important one. And I briefly mentioned before, and those and that, detail is integrations. You know that data, as I mentioned, data might live in your own tenant. It might be available in Masuni file storage. It might be in a different API service that you have access to. But where AI is really, really relevant and really powerful is to combine the data that can be combined from different data sources so that you can actually get most out of that data. And that's why integrations is super relevant here. Integrations basically allow the data to be on the same place so that AI can combine the data. And when the user requests or asks for some specific, data, then it is not only coming from the the data source, that they have access from before, but it's also inflated with data that might be relevant, and it can be combined from other different data sources. So when we are talking about, Copilot, we need to talk about integrations. That's why it's so important. And it's like that. It's when you have a proper integration strategy, when you understand your data, when you understand that it's not only, relevant to expose the right data to the user, but also expose it to AI and AI models so that AI can understand if the data can be correlated and can be inflated before it's returned back to the user. Because it's that that is actually going to create insights. You want the Copilot that will give you that superpower that allow you to make the right decision. And if you are x and if you get the the right, insight from Copilot, it will make your, decision much easier and, worse, and a smarter decision as well. So let's let's briefly touch about that. What does it mean integrations and how it's a good integration strategy? So in in SIGAO, we do have, been using this with, for the last, 10 years with our customers. Of course, integration started as a a typical point to point integration system a that needs to talk with system lead. But things with, with Copilot has have changed. So what what we normally do is before we, we enable Copilot in our customers, we do, a validation, and then we sit down with the customer to properly plan to have a Google integration strategy. So if we start on the if if we look at the Elon Musk rocket on the right hand side of the of the of the slide, you see that the first thing we typically do is we try to map out all the relevant data sources. This is this slide is made, of course, for the energy industry, but you can think about it as a, a general slide that can be applied to any industry. Identifying data sources and every every everyone needs to do it before we plan, and and have a good integration strategy. So everything needs to start by the way by that in in identifying the right data sources. Then as Jim mentioned, some of this data might be available for all the users. Some slice of the data might be only available for HR. Some slice of the data might be only available for finance. So the data assessment and classification is also a very critical step here. We know that, in in today's world, security is super relevant, and security starts by, only allowing the the data to be exposed to the users that do have access to it. So and that's super important and super relevant. In our energy industry where I work, this is even more, relevant these days as you can imagine because, security, is, is a concern, and we are super, sensitive when we have, to do this kind of data assessment and classification. And then, of course, depending on the cloud provider you, you are using, in this case, of course, we are using Azure and most of our customers and projects. We work with directly with Microsoft to make sure that the data lines in the in Azure with the proper security, with all the compliance regulation in place, and we validate that with both Microsoft and the customer. And then we move from, from, from there. So we build MVPs. We, design the the the solution, and we, we deliver it, and combined with the customer, we, lay down the road map, for the future. This can be applied to pretty much all the scenarios, and I think this is super important to Google your copilot discussions in your company, because beef be if you don't have a proper integration strategy, the copilot, will, not succeed. And that's what we see, in, some of the customers that have tried to onboard copilot experience without having a proper data strategy and integration strategy. And when you have this in place, then you are able to launch your copilots, because then you understand where the data is coming from, how safe or insecure the data is, who should have access to that data, is the data combined with other data sources, then you are really ready for copilot. But integrations can be complex. So if we look at, this is a blueprint from Microsoft. Integrations can be quite tricky, can be very complex. It can be highly, critical for our organization. So Microsoft included pretty much all the details around security, about, compliance, everything on this blueprint. However, we do have the experience to work with our customers. So we built, this blueprint in Seagull, which basically includes, all the services in Azure that 99% of the integrations actually requires. And by the way, I'm showing you a slide that is completely updated, and that's on purpose. These are integration. This is the integration blueprint that we used to use, 5 years ago. Otherwise, that's you see that, okay, we do have, the connection between external data sources outside that live outside Azure into Azure, and then we can pretty much send the data to the, target system or expose it to, users via Power Platform or Microsoft Graph. But we are moving. We are moving, and we are adjusting with this new Copilot era. So here's here's an example of what, things look like today. This is the latest version of our blueprints. We don't only, are supporting the integration of external data sources to your tenant. We are also using and and working with Nasuni to bring their own connectors to Azure. And why is this important? We know that some of the legacy data sources that you still have on your company will not have a graph connector, and it will be difficult to create one. The reason is there's multiple reasons. Structured data is one of them. The how difficult it is to build a connector to, all brand service. All all different, data source how how different the data sources can be, where they where they are located. So here's our some some here in this slide, you can see some of the different flavors. This is also important to say, that some of these data sources might be combined. As I mentioned before, Copilot the goal of Copilot is to bring you that superpower. And if we are able to bring first in the first stage, this data to Azure, second stage allow AI models to combine data from different data sources, so enrich the data in a way. And the last step is actually to expose it to Copilot. So this is, our typical scenario. This is what we're working now. In the bottom of this slide, you see fabricate, which is a new service for Microsoft. Integrations can be completely stateless. However, in this new world of Copiloting, sometimes it's worth identifying which slices of this data might be worth storing in fabric. And why is that? Because it's, it's much easier for copilot and AI models to, to combine data from different sources if the data is available locally and much faster to, to combine than just relying always in the external data sources. So there's a a lot of different flavors here. As Jim mentioned as well, we can Copilot can use, Graph connectors to easily access the data that exists in the data storage in external data storage. Sometimes we need to work a little bit more on these integrations, because the data might not be in a accepted format from Copilot. So we need to bring it first to Azure, put it in a format that can be accepted and and and read by Copilot before the user actually gets that, information. G mentioned, and promised that I would show customer cases. I don't have a lot to show, but I have at least one, which I think is super interesting to, to talk about. We had a customer in, in the energy sector that had a lot of, offshore facilities. And by offshore, I mean, the oil rig, for example. And, you know, these oil rigs have a lot of different sensors. When when they are, searching for oil or gas, they do have a lot of sensors that, look at the temperature, changes, pressure changes. So, not only for security, for the for the people working on the on that facility, but also for, reporting and if there is a need to change sensors, replace sensors. So what we did, we discussed, we we first build integration, between, those facilities and Azure, and that we use IoT app. As I said, it's just a flavor. There's a lot of different options to, bring the data into Azure. Another thing, as I mentioned before, is the data was not, of course, in the format, that was, by default accepted by the pilots. It was not a Word document. It was not XML. So, we had to work, to bring this data first measure, put it in a more familiar place. In this case, we use Cosmos DB. And in Cosmos, then we have the ability to basically create the right format so that they are models and copilot specifically could look at the database and understand the connection between dots. What's the end use of this case? So imagine the scenario. We have, we have operators and engineers that go from time to time, change the sensor in one of these, oil platforms or all these, on these, offshore facilities. When they are back to the office, they need to read they need to write the, a report about what was the change about, what was the data that they saw coming from the sensor, why was the sensor faulty before, And now it looks like right now. That was a painful process. The report took days to complete. And why is that? Because they need to look for the data, well, that the data from that specific sensor from that top shelf facility, And then I need to they need to put it in a table format and put it in a word document. I need to go, write about it. Why do we think that the sensor was faulty? And after that, after the the replace or after switching to a different sensor, they need to do the same exact same, process, get the new data from the sensor, put it on the data on the table format, and explain why this is new. By adding the data available now in Cosmos DB, so Copilot in Word was able to by the user is writing the sensor name in this oil platform, ID something. Copilot is able to identify that, was able to fetch the data from the Cosmos DB, is able to view the table, is able to write about the data all the data looks like, and the same for that is tensor. So the engineer that did this operation, started by using 3 2 to 3 days to write this report with this new copilot. And with this new superpower that we, gave to the the the engineer, he is able to, write us a report in 20 minutes. This is a good, showcase. This is the superpower that I've been mentioning since the first slide that we, we are granting to the users and that we see, that's going to be, a massive, improvement in today's, workplace. So, that's all for me as well. Thank you so much. Alright. Alright. Thank you, everyone, for join hope you found that, insightful, and you can get a better understanding of, how Nasuni and and Copilot can can benefit your, your organization. As a next step, like, the session has been recorded, so we'll share this out, with everyone that's registered, please feel free just to reach out if you have any other questions or would like more information either from Nasuni or from Sigal, we'd be happy to to share that with you. So thank you everyone again for your time. Really looking forward to staying connected. Yeah. Have a great day.