Transcript
And as part of that, we identified that there are organizations out there that we questioned that are using AI in some capacity. Now, AI could be using AI ticket resolution. So there was mostly that. There were some using it for intelligent escalation. We did have knowledge management and using AI for assisting with that, and also self-service experience. So there are organizations out there that are using AI and using it very well. So as we go through this, we'll sort of look at a bit more of what is AI, what are the use cases, and how you could benefit from the AI that we have within there. And we'll touch on a couple of future topics that are recurring. So our vision for AI within ITSM is really fairly straightforward. It's leverage AI and machine learning, and machine learning has been around for a while. Really generative AI has really transformed the industry, is provide you with a responsible, secure, and trusted way for relevant ITSM and ESM use cases to weave human intelligence and technology together to do great things. And I want to stress the wording there that we talk about, which is responsible, secure, and trusted, because that's a very important aspect of AI, whether we're using it in our own home lives, from a consumer perspective, or whether we use it in our business. We want to make sure that that data, we can trust this information, but more importantly, that data is secure. So what I want to say is one of the key things about this is we don't go out to the internet and go and look up things for you. Our AI is only within your Avanti instance, is only within your Avanti environment. We only look at your data. So whatever data is in there, we'll use from an AI perspective. So let's have a look at some use, couple of use, three use cases, and sort of walk through how AI can assist you and drive efficiencies within your organization. And in this particular use case, we're looking at it from an incident resolution perspective. So let's start in the top left-hand corner, where the end user has an issue. Let's say they've got a printer failure or something like that. Now, typically, they would get into the self-service portal or maybe a virtual agent, and they'd go and start searching knowledge bases or asking questions of the virtual agent of, I've had this issue, what can I do to resolve that? And they would go through a process of hopefully trying to resolve this so that your organization has a bit more of that shift-left approach. And if they can't resolve the issue, they'll go and raise an incident with the service desk. Now, one of the things that we do within that is use AI to generate what we call a ticket summarization. Now, that ticket summarization grabs what they've been looking at. So if they've looked at two or three different knowledge articles, that information will appear inside the incident. So when the service desk analyst gets that information, they already have that information in there. So they're aware of what that user has used from a knowledge management side of things or looked at. And then we automatically go through our machine learning and we classify the incident. So automatically classifying the incident means we get the information correctly stored. So from a reporting side, it's great. But more importantly, we can then assign it, make sure it's assigned to the correct team. So we've already used ticket summarization to get the information the end users looked at, and we've used machine learning to make sure we automatically classify the incident correctly. And now the service desk analyst is going to start working on the incident. Now, they might find that when they get into that incident, maybe other people have worked on the incident, or maybe there's a lot of information in there. And typically you would go through as a service desk analyst, going through all the various notes trying to work out what has happened within this incident. But using AI and generative AI, we can generate an incident summarization, which will give us the key information that's within that incident. So now the service desk analyst has got up to speed a lot quicker. They haven't had to read too many articles, they haven't had to go into too much depth. They're using the incident summarization to give them information that's very quickly obtainable and give them a quick way of getting started on that incident. And the whole concept around these so far is about reducing that mean time to repair. How can we quickly get to the point where we can resolve this issue? And then when we do resolve the issue, we want to start building out a better knowledge base. So we use generative AI again to generate knowledge articles. We generate a draft knowledge article where we can then go through the standard organizational process of reviewing it, approving it, and publish that article, where it can then be used by the end users or the analysts. And this then improves your user experience because now we've got better articles and we've got more articles out there. So the next time when an end user has an issue, maybe they can resolve this themselves. So that's using that from an incident resolution side of things. And we haven't touched on every AI use case at this point, but we're looking at, here's an example of where AI can help you. Now what about when we get to surveys? Surveys can become a little bit harder sometimes to understand what the results are. In a survey, you'll have a yes, no, or some sort of star rating, which is good information. It's quite definitive of what you're talking about. But when you add in free text and ask for an opinion, often that can be quite consuming to go and analyze all the results of that free text information across quite a lot of surveys. So again, this is where we can use generative AI, summarize all those free text results so that we get a summary analysis, which highlights keywords of this particular words that they're using there, and also obtain the sentiment within those text responses. So what is the person feeling about when they go through that process of the survey? So we start to build out some sentiment analysis, and we'll look at that in the demonstration during the session. But how can AI help you? How does it go about the process of helping you and your organization? It helps you reduce your service desk workload. So clearly see what the end user has already tried and looked at. If you're coming into the incident, whether it's halfway through or there's a lot of information, quickly understand the issues through incident summarization. Simplify your knowledge article creation, and in simplifying your knowledge article creation, you'll probably start generating a lot more knowledge articles and building out the content and the data that you have available. Having intelligent automation, so if there's an automation routine that can quickly be used to resolve an issue, you have that available to you. AI can help you enhance the self-service experience. The natural language virtual agent can enable an end user to easily find answers and interact with the user. Identify issues before the end user using proactive service management, and provide a better knowledge base so that information is there available to that end user. Gain faster insights. Again, using incident summarization, you can gain very quick insight into what's happening with the incident. And as an analyst or a team lead, you can start to create better dashboards, create dashboards against your important data quickly using natural language and generated AI to generate those dashboards. You can prevent your IT issues, so anomaly detections and proactive service management has the ability to identify and rectify issues before the user even knows, whether it's a device issue, an application issue, security, or performance degradation. And that helps you reduce your mean time to repair, you're identifying and rectifying issues before the user without IT intervention, or you're providing self-healing automation to speed up the resolution. And understanding the employee sentiment. We're all about the digital experience and understanding the user's digital experience. And part of that digital experience is that sentiment analysis. So sentiment is more than just a survey. It's a way of analyzing those survey results in the comments in the free text form that they put in there. And that will improve your digital experience because now you start to understand what the end user is experiencing. So enough of me talking for now, I want to run you through a demonstration. And in that demonstration, we're going to look at AI across virtual agent, self-healing, incident classification, incident summary, knowledge generation, dashboard generation, and sentiment analysis. So let's go and have a look at the demonstration now. So we're going to start looking at this from proactive service management. And it's the ability to be able to identify and rectify issues before the user's even aware of them. So in this particular scenario, let's say you're just working away on your laptop, your normal day-to-day workload, and then you get a message within Microsoft Teams that says you have an issue. So here we have, we have our proactive IT support team have come out and identified that my print spooler is not working. Now I haven't tried to print anything. So I'm not aware of that. Would they like me to restart the printer for them? And I'm going to say yes, I'd love that to be fixed. Now what happens is it will go through the process of checking again to make sure it's not working, restarting it, and making sure it's active. And then we'll get a response from the proactive IT support bot to say, we want you to go and check to see if your printer is working now. I'll go off and do my prints. It might have worked, it might have failed. Now if I put print failed, it will go and create an incident for me within ITSM. And then we can have someone work with that from the service desk. If I say the print works, then I can go and potentially just close off this particular proactive service management bot. Or I could go for reporting purposes, go and create an incident as resolved so that we're aware of what we've actually worked against. And that's proactive service management from the point of view of how that works within our organization. And these neurons bots are the main driver behind that. So we've looked at proactive service management and the ability to identify a problem before the user is aware of them. But what about the scenario where they have identified a problem and we want to do a bit more about shift left, get the user to be able to resolve the problem in their own right. So in this scenario, we've got self-service portal, which is a traditional way a user would come in to try and log an issue. They could search the knowledge base, but in this case, they've already identified the printer's not working. So they go in and want to log an incident, my printer is not working. And then to help the user be able to fix this in their own right, we come back automatically with a list of potential knowledge articles that might help them. So I can come through here, maybe it's something to do with how do I reboot or restart my network printer, and I can have the different steps there. I could resolve that in my own right, or what else do we have, how to add a network printer or a local printer. And again, we could put that and go through that process. And if we're unable to resolve that in our own right, we can then move to the next stage of this particular process, which would be to get what's happening with my printer and why is it not working. So that's one way we can actually interact through the self-service portal. So our other approach is to go out to our virtual agent and use natural language processing to look at the virtual agent. So we can click on the virtual agent, and that will bring up our capabilities to log into our virtual agent. So here we are, we've connected into the virtual agent. And I just want to put in there, I'm unable to print. And we'll go and see what the virtual agent can do for me. Now the virtual agent is much more interactive. So it wants to try and resolve this for you. So I'm going to say the printer is not connecting. And we'll see what our virtual agent can come back with from our point of view of how do I answer this particular question. Now the virtual agent, as I've talked about, much more interactive. What device do I have? I have Windows 11, so it's asked me about that. It's confirming that the problem is the printer is not connecting. And we now go through to our knowledge base to see what we have that might be able to help resolve this issue. And see a knowledge article that might be useful. We can copy that link. And we can then interact again, continuing on that by telling the virtual agent we're not connecting. And then we can go through different steps. So this guides the user much more through the problem. And in the end, we can log in an incident with the service desk to have that rectified as well. So now let's say that the incident has been logged with the service desk, so the printer is not working. So here we have our printer not working. I cannot print. We can see that our machine learning capabilities, which is our ticket classification, has automatically classified this as a printer failure. And the one other thing that we have is automatically we come back with options of how do we can fix this issue. So in this particular case, we're saying we can run a routine that's called an automation routine that is called check print spooler. So that will enable me to go and run that against the device that we would link to that particular issue. So I'm just going to link my device that's working for me. And I could go and execute that automatically against that device. So let's go and look at what would happen at the user's device to see how that works. Now I've said save and execute from the incident. So what will happen is that will initiate our Neurons Healing Bot and in a similar way to what we did with proactive service management, we'll have a Microsoft Teams conversation with the user to try and resolve this. And this is different to actually physically going a remote desktop connection into that user. So this could happen behind the scenes without any interaction, but this scenario we're interacting with the user. I want to fix it. Printer issues. Are you happy? Yes, we'll proceed with that. And again, it will go through the same process. Is it running? Is it not running? How do we resolve this particular issue? And maybe there's something else happening within that. So in a similar fashion that we had before, we're going to resolve this and automatically execute against that. So we can start to simplify and reduce the mean time to repair for the user. We've got some incidents here. Let's have a look at this one, which is the antivirus is not updating on my laptop. I'm going to drill into that particular incident, and we can see that there has been quite a bit of activity history through this. So normally what you'd be doing in catching up and understanding an incident is you'd be going through these different notes and trying to understand what's happened. Where are we at? What's going on? Now, what we do is click on the summarize incident, and what it's going to do is look at all the information that there is, and come back with a short summary of where we're at with this particular incident. Now I can see we've got our summary, which talks about what the laptop is, what the issue is, what the incident ID is, and then we've got our summary. We talk about what's happened, when did it start, what we went through to get this resolved. And then when it did reoccur, which occurred later, what have we done? How did we stop this from working again? And then here we've got all the different notes that we used as our data source. This is all the information we have to summarize our incident, and we can do that against any of our incidents. So we've looked at that one, and then what we'll do is we'll go and have a look at this laptop is always running very hot, and again we can do the same thing, summarize that incident, and get that quick update of where we are within the whole incident process. And as we can see, we've got our summary information again, what did we do, how did we resolve this, what did we need to get to, and what we use as our information. Now if we can put this into a resolve state, we then have the ability to generate a knowledge article. Now traditionally you'd go through, and you'd be looking at your resolution notes, and what you had to do. Now I can just say I want to generate a knowledge article, issue in resolution, and it will go and build that draft version of the knowledge article for us. And here we have that, we have our title, keywords that might be relevant to that. We can then put it against what particular category do we want that to go for, what the description was, so what was causing this, and what the resolution was. And then we can save that as a draft, and we can go and edit that. And that's our generative AI knowledge creation. Let's have a look at our generative AI capabilities within our dashboards. What we're able to do is either via predefined prompts, which we have here, or our own prompt, we can generate a dashboard widget without having to go into the dashboard editor. Let's see what we can get from the show all incident records grouped by status predefined prompt. Now let's return to our dashboard widget, which is as we requested it, and we can adjust that. We can change that to a pie chart. We can toggle the legends, we can maximize it, or we can export that image. So that's using our predefined prompts, but we can also type in our own prompt. And let's do show all problems grouped by status. And now we have that particular widget created for us. Again, we can change the types and the styles. And we can go and add that to a dashboard, or go and create a new dashboard. And that's our generative AI capabilities against our dashboards. What we're going to look at now is sentiment analysis within the Avanti Neuron Survey Bot. So the Avanti Neuron Survey Bot will send a survey to a user or a group of devices, and they'll respond to the question answers. And the typical scenario would be here where they ask questions around, how's the battery life? Have you had issues with it? What responses might you have in there? Hardware performance? What do you think might be needed for that? What's the overall support and maintenance? So all of these will get captured, and then we'll analyze that information when we look at the final results. So when we look at our survey results, we can see when the surveys were taken, what the responses were like. But more importantly, we start to summarize and use AI to give us a view of what the responses were. So those text responses that typically you wouldn't be able to capture without reading through it. And we can see here people frustrated about long startup and shutdown times, as well as running Microsoft Teams, common theme seems to be about RAM and CPU, and the sentiment was moderate and balanced. Now that's the overall view of all the questions, but we can also then drill into the specific questions and look at what were the responses within there. And then we can start to look at, have you experienced noticeable delays or performance issues? So we can start to summarize and use this information to get a view of how am I going? What's the core issues that are occurring? What are the survey respondents actually giving us information-wise? And that will be against any of the questions that have been entered. We can start to summarize those questions to start to get a view of what is causing the big issue. And that's the utilization of our AI capabilities within sentiment analysis for the Avanti Neuron survey. All right. So hopefully that demonstration gave you a good summarization or a good introduction to our AI capabilities within the Avanti Neuron solutions. And just to recap all that, what did we look at? So we looked at that incident summarization where a service desk analyst can easily understand what was happening within an incident and get that view of all the data and all the information that was there. We looked at knowledge generation, the ability to generate knowledge articles from within your incident and generate a draft knowledge article that you can then run through your normal knowledge management approval processes. We didn't look at ticket summarization in that particular scenario, but as I talked about, we go through that process of ticket summarization of what the end user has looked at. We looked at dashboard widget generation and the ability to use natural language input to generate a dashboard widget, which you can then add to your dashboards. We had a look at our virtual agent capabilities within the IT side of the virtual agent, but the virtual agent can also be used across HR facilities. Incident classification occurred underneath the covers through the machine learning. And just for a note, for those who have looked at machine learning before within the Avanti Neuron solution, we are re-architecting that within an upcoming release where before you were dependent on having the Avanti Neuron's platform to be able to do incident classification. What we're doing is re-architected that to be a standalone within the Avanti Neuron's ITSM solution. The digital employee experience score with the sentiments, it's survey sentiment. We saw how that works within there. We looked at self-healing and its ability to do proactive service management and also speed up through automation where it does go into their system. And incident correlation, I want to talk a little bit more about. So incident correlation we have had for a while, but we've re-architected our incident correlation. So in the 25.2 release round about April, incident correlation will be enhanced significantly. And let's have a look at what that means. So incident correlation is the ability to identify the commonalities between incidents to determine if they are related or part of a larger issue. So we're going to use AI to go and look at the incidents, look at whether there's some commonality or cluster of information and put them together so that from an incident side of things, you can look at, maybe I need to turn this into a master incident. It might be that this is a problem that we need to start putting through the problem process. So what we want to do is get that incident correlation there, get some similarity analysis and clustering and enable you to then be able to work out what's actually happening in my environment and why am I getting such commonality? So the aim is to reduce the disruption to your service and improve your employee communication by focusing on one incident and not a whole lot of other incidents that might be there. So as I said, we have had it for a while. We've completely revamped that, completely re-architecting that and it will be available in the 25.2 release. So what we aim to do and what we do from a roadmap perspective and what we are looking from an architectural side of things, we want to look at how we can elevate the user experience with AI, personalize and optimize the user experience. So using AI driven enhancements. So that could be within a chat type conversation. So we're looking at, do we put a chat GPT type interface on the front of that? So you'll ask a question that comes back with the relevant concepts. Using an AI co-pilot functionalities to assist and enhance efficiencies across any of the ITSM personas. Enhancing the automation through AI driven automation and you'll see that start to work and globalizing instantly. So looking at generative AI to automatically change it from the language that the incident is in to a language that you can understand. And developing that across self-service, the focus on scalability to make sure this works across no matter how many queries you're running against AI. Leverage predictive analytics, so implement AI driven predictive analytics for proactive management and to fully automate where we can, the enterprise service management utilizing AI. And I just want to reiterate what we're looking at here and what we drive is that responsible, secure, trusted AI. We only look at the data within your Ivanti instances. We don't look at another instance data. We don't go out to the internet and go and grab information from the larger internet environments. We're only using data that's within your environment because that we feel is the best way to trust the information you've got because it's your information. We feel that's the most responsible way of doing it. And certainly we feel it's the most secure way of looking at that data. So I know there are a few questions, so I want to sort of go through and look at the questions. There was a question, is this going to be available for SAS customers only? Yes, it is SAS customers only. We have no plans to put that on-prem. The architecture underneath uses open AI within our architecture. So we've installed it within our own environment. So we do run that as a SAS only option. From an AI capabilities looking at for the future, we sort of touched on a few of those. Translation is one of the concepts. Translation looking at how do we translate information dynamically for you. We are looking at automation capabilities from an admin side of things. Sometimes we lose what an admin can do, don't get talked about from the AI side, but there's a few options there. Generating it, if you've ever written expressions within as an admin of ITCM, you'll see that expressions can be sometimes complex and hard to understand. We're going to use generative AI to help you generate the expressions within the expression builder. So you can type in what you want, it will build it for you. And using AI within the admin side as well to potentially build new business objects and parts of the organization as well. So we are looking at that. And not surprisingly, you'll hear a lot around agentic AI or agent AI. Fairly new concept. I know that Salesforce were advertising that fairly big during the Super Bowl. And it's something that's developing at this point in time. So it's not a core capability we have in the product at the moment. We have the bones of that. If you look at how agentic AI lays together, it's really the intelligence on the front end for the gen AI that goes and calls workflows that you've already defined. We already have all the automation workflows under the covers. So we're just looking at how do we put those agent AIs in front of that to orchestrate the automation of that. And that is on this. We're looking at how we implement that within the system. So everything that we talked about to the tools and features you're presenting are including the current version neurons. As I said, incident correlation is getting enhanced and the new version of that is 2025.2, which is the April release. Everything else I've shown you is available now. You do need to upgrade from a licensing side of things. I know there's another question there. Is there a licensing requirement for that? You do need to get a license to get the capabilities. If you're a standalone ITSM customer, there is a licensing option to add AI to that. If you're one of the solution bundle, use purchase licenses, then it's the premium license or premium enterprise license that you will require. It's the Avanti Neurons AI for other spaces such as AI. Not yet, but we are looking at what are the use cases across ITAM, for example, and also looking at bringing AI across the enterprise service management side of things. So HR facilities, GRC, and other areas. I have just said there is an added cost to add AI to your environments. AI capabilities available for service request. Again, we're looking at what the use case is for that. Happy to listen to any use cases you might have, but we're looking at what that might mean for us. A few more questions about additional costs on those, so I think I've answered those. Is AI going to be implemented automatically or will it need to be enabled? Once you've got a license, it's enabled within your instance. And then within the Avanti Neurons instance, there is an AI configuration hub. And that AI configuration hub then enables you to turn on or off functionality within AI. So let's say, for example, you don't want to use generative AI for knowledge management yet. Not something that you're quite ready for. You can toggle that off, but you can still have incident summarization, ticket summarization, machine learning, all the different other components that are in there. And then when you are ready, you can turn that on. You also have the option to modify some of those capabilities. So when you look at something like incident classification or knowledge generation, it looks at certain fields to go and generate its information. If you have additional fields that you want to add into that, you can extend that and include those. And it's all just putting in and configuring those. So you've got the ability to have some flexibility. So that's all the questions. We're just a couple of minutes over time, but that's fine. I wanted to make sure we got all the questions answered. Thank you very much for joining us. I hope we've gained a lot of information around what our AI capabilities are. We are developing them, and then we'll be road map sessions over the year, I'm sure, when our R&D team talk about what's next within that. And I hope you've got a lot of information out of this, and thank you very much for joining us today.
