GenAI in the Real World: A Customer’s Perspective at SKILup Days 2024

Hi everybody. Um, I am, uh, ademi and, uh, in the coming, um, 30 minutes we're gonna share about generative AI in the real world, uh, and a customer perspective. And with me is Andre, uh, who is one of sis eight clients that jumped on the AI wagon, and he's gonna share his perspective, uh, around the journey.

Okay. Sounds good. Sounds good.

Um, so maybe, uh, Andre, do you wanna say a few words, um, about yourself? Um, well, I've been an IT manager for about 20 years. Um, I've covered the range of technologies coming up from when, um, AI was, uh, not in somebody's eye.

And at this point in time, I'm jumping on the bandwagon because I see the benefit it has in terms of helping users get what they need faster and helping me to get what I need to them quicker as well to Great. So I'm Caid. CEO Caid is about, uh, I'll just move on.

I'll share a bit. Caid is about, uh, 20 plus years, uh, company. We have thousands of, uh, customers across Halton, 40 countries in 42 languages.

And we provide, um, it service management, um, uh, software, uh, for a variety of, of segments, uh, uh, in the world. Um, I've joined Caid about a year, getting close to a year and a half, um, after, uh, uh, a personal journey through, uh, my own startup and then being part of a big, uh, public, uh, American based, uh, public company. Um, so excited to be here and, and share with you, um, what do we do and what is generative ai, uh, and in the world of ITSM, uh, maybe Andre a bit about, uh, iTech.

Uh, well, iTech, as you can see, is based in Jamaica. Um, they have a couple of locations across the island and they work on drip irrigation solutions, solar energy solutions, um, um, water-based solutions. And, um, we have basically done a lot in terms of helping the development of the country for all the farmers that we have here.

For all of the people who are looking to get start in renewable energy, as well as, um, what we partner with a couple other Caribbean countries to extend the same facilities there. Nice evolving, uh, technologies around, um, everything where the world is going, uh, for water, for electricity. I, I think it's, uh, it's doing good to the world, that's for sure.

So absolutely. What we gonna talk about today is we're, we're gonna take two trends we see from all thousands of clients. One is the rise of ai, self-service, and the second one is the automation of itil, uh, with ai.

And we'll try to explain a bit about them, share some data, and also see live demo of how this is used within Isla tech environment. Okay. So what you're seeing here is some sort of the data we we got with serving, uh, a few months ago, thousands of our clients, um, and how they see, um, the thing that they're focused on.

So any one of our customers enjoy a better employee experience thanks to AI tools. Um, and it means that almost 28% of them aim to improve employee experience. So through an increased self-service is your resolution, right?

And when you think about it, people are texting, people are messaging in this session, in these 30 minutes, everybody will go, will text. It doesn't matter, even if their eyes are in front of the computer, even here and Andre Mm-hmm, while we talk, we will keep texting and doing things here you can see on phone, that's the world where we live in. And when our employees, they need help, they need to resolve something that they expect an experience that more conversational as part of what how they like is done today.

And I wanted to ask, um, ask you, Andre, how did, how this, this came across in your org? Maybe say a few words. What, what do you see your company, is it the same, like the what we're talking here or, uh, I don't know, share a bit?

What do, what do you hear from your employees? Well, um, I'm basically pushing the adoption of the AI self-service, um, solution here. Um, it was something I started to see in other services and when I came across the sales offering, I was actually pleasantly surprised because my past experience with the help desk solution involved things along the lines of having to do all of the heavy lifting myself or have a clerk take care of it.

So, for example, categorization was a nasty bear that had to be dealt with and was not always consistent in terms of what people put together in terms of categories. Of course, that's a simple automation task now, but, um, with other sys, with other help desk solutions, not so much. So it, you can imagine that that takes a lot of time in terms of just trying to support people.

Um, in terms of the self-service part of ai, I am seeing a steady, but a gradual increase in the use of it. It is something that you can send an email and, and, and send out, um, what do you call it, um, group WhatsApp and say, Hey guys, there's this wonderful, um, application called CE that you can literally just start to talk to and it will tell you how to do X, Y, and Z whenever you're having problems. Um, it's not so easy to adopt, but having the ability to do it makes it a lot easier to, well, let's see it, have some sort of management, Let's see it together.

So, um, before we come to the benefits, I will present, uh, my screen. Like many other companies have their own, um, uh, through cs a their own, uh, service portal, right? Which is like FAQs, um, open tickets in this case, I don't have any open tickets and more and more areas like password reset and, and, and requesting different things.

And with that, the self-service, AI is an experience that they have the option you see here on the left and here, well, the self-service portal transformed to be a conversational experience. Andre, give me a a good question. You think it's worth to ask the AI that your employees would ask?

Simple. A very simple thing would be, my printer is not working, how can I fix it? Now, a lot of time when people see the world of conversational, they just think that, hey, you probably go to the web and you find an answer or something like that.

But the way, uh, uh, we believe AI is the right way to do it, and you can see the answer here of how to solve it, is in the system, Andre has the opportunity to understand why this was the answer, right? And if we see here the question, right? My printer is not working, what you can see below is that there were service records in the past, whether it's in the last day, when the last week, or in the last month, that they were the one that used in order to generate the answer, right?

For example, user is unable to print and is requesting something user is experiencing problem with printing some knowledge based item around it. So the answer we're seeing here is an answer that is generated with the use of generative ai, that the system also makes sure that the fresh data is used into the orchestration of the right answer. Now, question for you, Andrea.

What happens today if someone needs to open a ticket or open an Info? Well, no, well, typically what happens is people love to chat. They like to use WhatsApp to send in information.

So instead of having them use WhatsApp, what I try to do is just tell them, send an email, please. It'll work. It'll get through to you.

You will actually get an immediate response. And one of the benefits of caid is that that immediate response is not restricted to working hours only. So it means that when our factory workers come in at seven o'clock in the morning and they have a problem, they can actually talk to somebody as in the AI and get responses that make sense.

One of the benefits of this, um, solution is that we can actually fine tune the responses that they get by putting in articles in the data pool and tell it, okay, if you need to add a printer to where you are, here is what you can click on. And immediately, instead of somebody coming to install a printer for you, they get a link, they click the link, the printer is installed, that saves time, that provides solutions that are right there and then, and it's a 24 7 thing. You don't have to be there to take care of people.

Yeah. So you are saying for your place, 24 7 has a big meaning, and also you are using the data pool to keep adding information that is not only in the tickets, et cetera. So one way is by the email that the AI knows how to answer, but also, like you said, people in the world of today, they don't wanna fill out forms, right?

They don't feel comfortable. You know, I say some of the young people, they don't even know what the word hardware, right? Um, or something, some things like us, uh, you know, that, that maybe 20 years back knew that these words.

So the way it works is that, let's say this didn't resolve my problem with the printer, right? So instead of filling out the form, or even in this case going and sending an email and explaining my issue, The User just clicks Click boom, boom. Yep.

So What's what happens now, Andre? What do, what's going happen? What's happening now?

Well, at this point in time, once a ticket is created, we literally are alerted to the fact that there's a new ticket, we take a look at it and we move forward. The ability of the system though is that, um, I, I'm not sure if you want to highlight this, but we also have ai not just for users, but also for agents now. So when we have solutions that make sense, it literally becomes a quick chat with the system to provide the solutions that are full.

Okay. Yeah. I'll show it in a second.

So in this moment when this user said, I need help from a human mm-Hmm, the ticket is created, the AI sets the title, the description, the urgencies, the right categories as Andrea explained before, and everything is ready for the agent in this case to start working on it. And for the user experience. Mm-Hmm.

It's conversational. So it's twenty four seven conversational and no phone filling. And now on the left side it has like unresolved.

So until this issue is resolved, this stays as it is, uh, waiting for him. So we talked about the benefits, right? Of this AI self service.

I think it's also an opportunity Andre Shield, like how people are using it. Um, today, if we move on, um, you know, I'm curious for you Andre. Um, people when they think about ai, they, they always, like, they would say, wow, it's hard to implement.

Uh, wow, there's probably some challenges around security, and how do I make sure that the AI is not using data from other conversations of a different user with another ticket That was, it's a personal information. So maybe share a bit, what's your experience on it? Well, my experience with onboarding, um, or bringing AI to the, to Israel tech has been actually one where it was quick.

It was easy. It was, it allowed me to clearly identify guardrails that people could or could not cross. It allows me to, um, depend, sorry, separate documents that can be sent to everybody, or content that can be sent to everybody versus content that cannot.

And one of the easy things to do was to upload our ethics document or HR documents so that questions that people have normally as simple as a new user coming on and saying, Hey, how long is lunch? Or when can I go? Um, yeah, can get answered right away conversationally immediately.

And that was a big game changer. So in one or one, let's call, I'm gonna call it six weeks. In about six weeks, we were actually able to push this out and say, here you go, this is functional, it's done.

And it took that long because I didn't have all the time that caid was, was, was us was trying to get me to invest in the project. So six weeks is actually very, very, very long. Think of it more like half the time, You know, I, I think the biggest thing that happens with these AI projects that, um, is that once you see it working, it change your mind to everything else you're doing.

Not only to it, right? It just frame your thought or anything I'm doing in my life, AI now can take them better or do instead of me or reframe my thinking around it. Okay?

The second trend is automation of ITIL with ai. So ITIL for those on the call that are IT manager, it's the right, the, there's a lot of certificates around it where problem management, incident management, knowledge management, change management, how you do them, what's the right protocols to do that. But some of them, mm-hmm.

If you go one by one through these IT shops and you see that it's very hard to do it. It takes a lot of time to go every day and, um, find 10 incidents that are correlated to one problem and create that problem and connect the incident and or when you're doing a change to really detail all the risk associated before the change is, is, is going out, right? So, um, what we wanna show you here and Andrea will show is how Isla is using AI to make the it work much better.

What you're seeing here, maybe you wanna share a few words. What are they seeing here? Andrea, this is your queue.

Okay? Um, this queue is actually, um, a lot is a lot more flexible than other options that I've used in the past. Something here that, um, you might want to highlight is when you're going over the tickets, you can quickly click on the actual items and change priorities change.

Um, you can change pretty much everything that's on the grid, hovering over the ID or the description allows AI to literally summarize what the ticket is so you can quickly see where it is, what's been happening, what needs to be done next. And not only that, you can always see when you are in the red when bad things are happening for your company. Um, yeah, I can understand from what you're looking at right now, you can quickly identify that we have some old tickets that are there and that's simply because they keep coming.

They just don't stop. Okay? Um, if you keep, if you look along the lines, we made a change recently to actually make sure it's being sent to the right.

Um, so I should say, sorry, this is actually a very flexible system. You can actually automatically categorize things. You can take AI to AI to these categories.

If you see those stars beside the, um, right where you're right, you are, Yeah. That AI say maybe you can have a fine tuned categories Yep. Based on the information that's in the ticket.

And I'll let you know that typically once categories are set in other help desk systems, nobody goes back and takes a look at them. And as a result of that, the reports that you get don't make as much sense as they should. They don't give you the good, the big picture.

It. So we talked about ITIL also, right? So when you click here, just like the end user has a conversational experience also here, the agent, the admin, he can connect to, uh, uh, uh, a chat bot that helps them do their day to day and especially the itil.

That's very tough. So Andrea, let's look at something on your system, for example. Uh, uh, what tickets have corelation to, um, correlation.

Let me correlations, correlations to similar topics aloud topics. Right now, all you is available, right? Take a look what's happening.

In a glance of two, three seconds, you are able to see you have two tickets around purchasing three tickets here around user workstation, category three tickets around this one, et cetera. And you are able now to work with it to really conversationally ask, what did you tell me, Andre, how do you use it? What kind of question?

I, I remember you gave me some good example what to do next, right? Yes, exactly. What's my next highest, highest priority ticket?

My next highest your ticket. I'll leave the typo. Usually the AI knows how to fix it also, It found right one, and it also redirected me to this one, right?

Okay. So in this case, we're gonna do something else. We can ask here and Andrea, right?

Which is like, what do you suggest for me to do here? Let's see if the AI looking at all the tickets in the last few years, right? And using all of the knowledge base and everything in the system, it gives you also a suggestion how to solve it based on the details provided in record 72, which is this one, it seemed you are tasked with proceeding the, with the order for the Dell server number two, and which here are the steps you should do.

Now what we're gonna show here is that also this is monitored. So you see what are you suggest me to do here? And he, these are the service record that were used to generate the answer, Right?

Right? Yep, That's right. That's amazing stuff.

How So? For admins that don't know what to do, this is actually something that's great. It'll help guide them as to what actions they should take and provide them with examples of what has been done in the past as well.

Yeah. I'll show another example. I know that we've talked about it, Andre, right?

Can you craft a message to the user? So AI is in a level where it's not only able to explain to you what to do, right? It, it's also providing you a pre-made email to send, right?

Please confirm if you'd like to send this message confirmed and you what you would see in this case that the AI also you see on the left side, put it in in a way that it's formatted and everything. Of course I wanna do cancel so it doesn't send it this regard. What else?

You played with it and saw in it. Um, that's interesting. Uh, Andre, what, what, what kind of more things you want us to show here?

Well, one of the things I I was looking at was summarizing, um, the problems. Let's say in the first 50 tickets, I wanted to see if there were any, um, issues that were the same. Because typically when you have a situation, you're gonna end up with like five or six people sending you the same situ, same problem In a typical help desk.

You, you have to go look at each one to see what's, what's related when you do this, when you look at this and you say, okay, in the first 50 tickets, how many of these tickets are the same? Yeah, you can quickly, you can quickly identify where you have things that need to be merged. And if, um, in a future iteration, I'm expecting that Caid is actually going to be able to merge tickets by using the chat bot.

Um, yeah, I'd expect that in less than a year. Yes. Yes.

Um, in this case, it was asking, do you have tickets that need to be merged? Probably for those mm-Hmm, that are open. It says it, it didn't fine.

Um, but you, you are right. The first phase is to take the data and the AI helping you to resolve, helping you to know what's the right thing from the vast of few years of data or even two weeks. And everything is refreshed in real time all the time.

The next phase of AI is a lot of automations, right? So if you wanna fix something, you can have the AI offer you shell scripts and also offer you to execute them if you approve it. Because ITSM solution also have the agent within the laptops or the servers.

And this will allow also to execute if the agent, uh, decide, uh, to do so. I was actually gonna suggest, um, one of the things that I did in the AI chat bot was ask the simple question, what can you do for me? And it actually generated a list of actions that I could take.

So that, so when you're an IT person, you don't know what the chat can support asking. That actually led to I yeah. Led to here Do great.

Wait, wait, I'm gonna, I'm gonna do it. I, I don't think I ever asked that question, so let's do it together. Okay, what do you do?

You ask like, what can you do for me? Yeah, simple. Yep, there you go.

Literally, nice, nice scheduling, troubleshooting, writing, creating everything was there. And it actually allowed you to drill down deeper into what it could actually do for you. So, um, going, uh, forward, um, we talked about and we saw how AI is transforming it.

Um, I, I wanna say thank you to Andre. Andre has been a valid client, uh, uh, uh, a, a great client for, for CIS aid and also he's, he likes to jump on innovation and a lot of great things. So I wanna say thank you to Andre.

Um, we have a lot of things cooking in, in the coming few months for more and more capabilities around ai. So I hope we'll keep, uh, serving you well, Andre. Yep.

There's a QR code, whoever wants to come play. We have like workshops where you can register and just on your own will give you access to the system, explain to you how to do things, and you can play with it and see the power of it. Um, I'll finish up in a world where I joined CS eight, about a year and a half ago because I believed that the revolution around generative AI and ITSM is gonna drive a, a almost fully AI contained environment for it.

Uh, not now, but, but it started. Companies are seeing 30, 40% already of the volume that's contained by ai and this number will only go up. Um, and, um, you know, if you're curious and if you want to touch it, if you wanna ask your own play with the questions, just take the QR code and, and be happy to show you how it works and give you access.

And again, thank you Andre, for the time.