Transcript
be looking at boosting service efficiency with AI digital transformation. Firstly, we'll look at a couple of concepts that support this. So pre-ticket automation. What is this? This is about being proactive. We have a series of healing bots within Neurons and you can build your own as well. And they allow us to go off and be proactive, go off and monitor devices and make sure that they are performing properly, look for any issues and if we find an issue, we'll attempt to resolve it. And if we can't resolve it, we hand it across to the service desk. This really helps the experience to the end user because often they won't log a ticket, they'll just bear with it. It could be a blue screen at their fed application that's performing poorly or now and again errors pop up and the end user just gets on with it and there is something in the region around we think 50% of incidents and they're actually logged, we just deal with it. So the idea about being proactive is we are detecting and fixing problems before they become incidents. We also have post-ticket automation. What we're doing here is we're putting the healing bots, these little agents in the hands of the service desk and even the end user, which is coming soon and allows the service desk to start to resolve issues on first call without having to actually escalate into second and third line or get more advice or do extra diagnostics. They don't have to call up the end user and do remote controls, which can be quite annoying for the end user. They're trying to get on with their own job and they're being impacted by the service desk who are trying to do their job by diagnosing and remediating the issue. So if they can have a bot to call upon, just simply press a button, maybe do all the diagnostics, potentially do a fix or remediation. Again it's a lot better experience for the end user. They don't have to be bothered. And again, this is we're operationalizing our existing knowledge. They could be work instructions, they could be scripts, they could be knowledge articles helping to resolve incidents on first call. So there are two concepts we're going to look at in the demo today. However, one thing I like to do is dumb down these concepts and something that's very, very consumable, something that's easy to understand. And I do this with dominions. I see the service desk analysts, the desktop engineers, the incident managers, etc. as grew and they want to remediate, they want to diagnose the mundane tasks that happen over and over again. The incidents are always coming. We want to use the minions, the bots to go and mediate them. So that's an easy way we can explain that to our peers, to our management, to people who are interested in proactive resolutions and these concepts of pre and post automation. And so examples here, we can have proactive minions going off and diagnosing issues on the desktops and the devices, trying to fix them and maybe hand them across to the service desk if they can. And then once the service desk get these tickets, they can use the bots to go off and attempt to fix or to diagnose the issue. So this is a very simple way we can explain these bots. And then when we compare it with an approach that we're used to, this reactive approach where we wait for someone to log a ticket, and once that ticket is logged, we start to work on it. Rather than waiting and hoping that end users log tickets for us, we're going to use the neurons bots to detect the issues, attempt to remediate. If they can't remediate, they log a high context ticket with the service desk and the service desk can help them quickly remediate that issue. The big difference is here is we don't have to wait for the employee to do something. Even using the most sophisticated AI chatbots and digital assistants, et cetera, it still requires the end user to make the effort to log a ticket, to talk to an AI bot, and ask for some help. But the big difference is the big shift in what we're doing here is the user doesn't have to do anything, and they may not even know they have an issue. However, we're looking to fix that for them. And then if we lift these concepts up into major incidents, large problems in the organization that are widespread, looking at our incident correlation capability, which looks at similar incidents coming into the service desk. This could be through the bots themselves logging the incidents on behalf of end users. It could be end users logging the issues. It could be a third party event management system pushing in events, could come through a whole myriad of sources. However, the incident correlation capability is analyzing all these events and incidents and creating what we call a cluster. And we'll have a look at that in the demonstration as well. So what's with that? I have a big list of tickets in my system here, and I can see there's a common theme of slow lock on performance. Various people have logged it. Again, it could come through the neurons bots, could be logged on behalf of these end users. These end users could be logging through the portal, through email that you may be calling the service desk. There could be third parties pushing information into the solution as well. However, we can see there's a lot of slow lock on performance tickets being logged. If I go to my incident correlation dashboard, I can see here that there are a number of clusters that have been logged, and this is kind of gathering similar incidents and similar tickets and putting them together into one single cluster. I can see here we've got 21 reported incidents of slow log on performance. So it spiked earlier on today, and that's started to slow down. These lines will elevate and spike and go down depending on the number of tickets being logged. And also the red and green will indicate that the rates of tickets being logged as well. Red indicating an increase in tickets, green showing that it's starting to slow down. And I can look at this log on performance cluster, and I can see all the tickets that have been analyzed and added to this cluster. At the moment, there isn't really any service management process built into this. It's just literally AI saying, here's things that look similar. However, now I can invoke some of my service management practices where I assign a parent, and this now marks this cluster as in progress. However, I can grab all the other tickets and link them to that parent. So now we have a child-parent relationship, and I can just use this parent as a single ticket. I'm going to start to diagnose and remediate, and obviously anything I do in this ticket will propagate down to these child entities and update the relevant end users. So I take a ticket that I'm going to start to work on. It's being logged. I can see it's coming through the Neurons bot, and it's being assigned to me, and I'm going to start working on this ticket. The bot also indicated which machine was at fault or had the issue, so I can have that linked as well automatically. I can use this then to start diagnosing the issue. I'm going to have a look at this CI. It has all the information pre-populated, plus all the discovery information that's coming through from Neurons as well, so I can get a snapshot of its current status. But more helpful to me is I can then open up that device in Neurons. So this is where we have done all the discovery, and all the information pretty much in real time is available to us in this view here. I launch it into the digital experience screen, and I can see everything seems to be okay from this perspective, so I haven't actually logged any issues yet around potential problems with the device. I'm going to continue investigating the issues with this device. Obviously with this discovery, I'm getting all the information here. I can see the performance, network information, CPU storage, et cetera, and a whole myriad of information as well. On the left-hand side, including all the granular details we can look at, and the exposures, the event history, what patches are being deployed to this particular device, the software on it, and also using edge intelligence to get some real-time information. Then on the right-hand side, again, this is where we talk about post-ticket automation, where I'm now investigating this issue that's been logged, and I can see here I've got some custom actions that I can use to help diagnose or remediate this issue. First of all, I'm going to hit the device diagnostics bot, and that's going to run directly against the agent on the machine that's running, pull back all sorts of bits of information that may be useful as part of my remediation. What I can see here, the last thing it's come up with is lock-on performance is classed as terrible. That was something that was logged previously, so that's in alignment to what has been logged by the end-users. I can also see here memory usage, disk queue, battery status, any CPU, et cetera, that's being consumed. All that information is in here as well. Then there's a myriad of other bots that I can use to do different things. Again, this is just a sample of what's in our templates and a sample of what you can build yourself. I don't think there's an issue with this endpoint. I'm going to now take it a step up and look across my entire fleet of devices and see if there's something specific that's going across everywhere. I'm going to use what we call Edge Intelligence. Edge Intelligence will query in real-time all the devices that are currently switched on and then talking back to Neuron's cloud. My question is, I just used this earlier so it remembers questions, but again, you can use all sorts of different natural language to query. I'm just going to ask to show me the lock-on performance for all devices, and that'll quickly come up with a dashboard in the red, amber, green to show us what's running well and what hasn't been. I can see here there's two users that have had a terrible experience with their login site. I can click into that. Looking at this, I can see two different users, two different machines have had this problem. This particular machine has recently had two sessions that have been classed as terrible, so I can click onto that. I can review the sessions, so I'll click on here. This is now going into the granular login script of that machine at that particular point in time. I've gone from a very high level looking at every machine in our fleet to drilling down to a very specific part of the login script, which is taking all the time, which is this. This is taking a minute. This is obviously where the problem is, and I can take this information back to the desktop engineers and say, hey, we've noticed that something is running against the logon script that's taken a long time. We can kind of remediate that, and it could be there's a change going on over the weekend. Maybe not tested properly. Maybe it's only been tested on a subset of devices, and it's causing now an impact to other devices as well. I can even click into this, and it'll give us a list view, and again, this is the problem child, and we can investigate which other devices are suffering from this. So again, we can kind of get a, if we get the right information here, we can analyze, is it a certain subset of operating systems? Is it a certain location? We can play around with the information here, slice and dice it to get some meaningful information back. So just to summarize what we saw today, we talked about, at the end, we realized there was an updated logon script that got rolled out. So this happened maybe over the weekend. It caused a poor experience for the end users. We saw some tickets being generated by bots, so also they could have been coming in through events and created manually through the portal, the service desk agents, create them through phone calls or even email, they could have come in that way. The service desk analysts reviewed the instant ticket, then we reviewed the correlation dashboard to see if there was a major issue. We then went on to investigate the issue and went onto the individual laptop itself in Workspace to actually then analyze individual components of that laptop. We then queried the entire fleet of laptops and then identified there was a bug in the logon script, which then we could pass on to whoever was responsible for that, the desktop engineers or the people developing the policies, and they could go and remediate that issue. So what we used today was Neuron's Workspace, Neuron's Healing, Neuron's ITSEM in combination, so they are then native integrations. So thanks for that. Hope that was useful, and we'll see you in the next session. Bye-bye.
